2018 Information and Communications for Development 2018 Information and Communications for Development Data-Driven Development © 2019 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved. 1 2 3 4 21 20 19 18 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immun- ities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creative- commons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2019. Information and Communications for Development 2018: Data-Driven Development. Information and Communications for Development. Washington, DC: World Bank. doi:10.1596/978-1-4648-1325-2. License: Creative Commons Attribution CC BY 3.0 IGO. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribu- tion: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. ISBN (paper): 978-1-4648-1325-2 ISBN (electronic): 978-1-4648-1355-9 DOI: 10.1596/978-1-4648-1325-2 Cover image: © Sam Smouha. Used with the permission of Sam Smouha. Cover design: Debra Naylor, Naylor Design. Library of Congress Cataloging-in-Publication Data has been requested. Contents Foreword ix Acknowledgments xi Abbreviations xiii Executive Summary xv Data deluge xv Who benefits? xv Data-driven business models xvii Data belongs to all of us xvii Notes xx References xxi Chapter 1 Data: The Fuel of the Future 1 Data, data, everywhere 1 How data is changing development 1 A data typology 2 How governments use data 5 Structure of the report 7 Notes 7 References 7 Chapter 2 Supply: Data Connectivity and Capacity 9 The ever-expanding data universe 9 Goodbye data carriers, hello data creators 10 Cloud computing: Back to the future 14 Internet of Things: Data is all around 17 Data-driven business models 18 Data holes: Filling the gaps 25 Conclusions: Toward sustainable national data ecosystems 27 Notes 29 References 30 v Chapter 3 Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 33 Introduction 33 The big data revolution 33 The evolution of artificial intelligence 34 Using big data and AI as a force for social good 35 From design to responsible use: Ethical challenges with using big data and AI 44 A way forward: Harnessing big data and AI to “leave no one behind” 47 Notes 48 References 48 Chapter 4 People and Data 51 Introduction 51 The data market 52 The benefits, costs, and risks for people 55 Remedies 59 Toward a more balanced data market 62 Looking to the future 63 Notes 64 References 65 Chapter 5 Firms and Data 69 Introduction 69 Digital platforms 69 Business models for digital platforms 71 Digital platform dynamics 73 Firms in the data economy 74 SMEs in the data economy 77 Looking ahead 83 Notes 84 References 84 Chapter 6 Policies for the Data Economy 89 Introduction 89 Policies for building data as an infrastructure asset 90 Data policies for building trust 93 Data security 99 Policies for maximizing the data economy 100 Notes 103 References 104 Data Notes Data for Development Indicators 107 Availability and users 107 Affordability and usage 107 Government 113 Infrastructure 113 Digital Adoption Index 124 vi Contents Notes 131 References 131 Bibliography 133 Contributors 149 BOXES 1.1 Open data tools for improving transport through big data 4 2.1 Sub-Saharan Africa: Reliable electricity and the digital economy 15 3.1 Using big data to predict dengue fever outbreaks in Pakistan 34 3.2 Artificial intelligence and the transport sector 35 3.3 Using machine learning to analyze radio broadcasts in Uganda 40 3.4 Estimating population counts and poverty in Afghanistan and Sudan 41 3.5 Mapping energy access in India 41 3.6 Cleaning Mexico City’s air with big data and climate policy 42 3.7 Self-driving cars 43 3.8 Monitoring public sentiment about policy reforms using social media in El Salvador 44 3.9 Shedding light on migration patterns using social media information 44 3.10 Data privacy, ethics, and protection: A guidance note on big data for achievement of the 2030 Agenda 46 4.1 Income-generating opportunities 57 5.1 Jumia: “Cash on delivery” e-commerce in Africa 72 5.2 Agribusiness SMEs and data-driven supply chains 79 5.3 Alibaba’s success: SMEs as the foundations of the business model 80 5.4 The app economy in the Arab world 81 5.5 Open data for SMEs: The European Union and Colombia 82 6.1 Defining a policy framework for open data: Mexico’s experience 92 FIGURES ES.1 The growing internet xvi ES.2 Types of personal data xvi ES.3 The Information and Communications for Development series xvii ES.4 Are you willing to share your data? xviii ES.5 Toward a new value chain for personal data xix B1.1.1 How open data tools can assist transport 4 2.1 Internet users and broadband speeds 10 2.2 Global IP traffic and global consumer IP traffic 11 2.3 Hyperscale data centers 13 2.4 Price of electricity (US cents per kilowatt-hour) 14 2.5 Global machine-to-machine connections and traffic 17 2.6 Machine-to-machine connections per 100 people, OECD member countries, June 2016 18 2.7 Global network traffic and retail telecom revenue, selected countries 20 2.8 Global advertising revenue 21 2.9 International carrier and over-the-top traffic (billions of minutes) 23 2.10 Mobile messages, year-over-year change (percent), 2014–15 23 2.11 Global internet protocol traffic and websites by language 26 Contents vii 2.12 Mobile data usage 27 2.13 Mobile data pricing 28 3.1 The Sustainable Development Goals 36 4.1 Types of personal data 52 4.2 The personal data market 55 5.1 Physical and virtual enablers 70 5.2 Market and behavioral enablers 70 5.3 Geographical concentration of digital multinational enterprises with revenue in excess of US$1 billion, by region, 2016 71 5.4 A methodological approach to assessing digital platforms in emerging markets 75 B5.2.1 How more data contributes to current business models in the food chain 79 B5.3.1 Alibaba’s physical and virtual enablers 80 6.1 A framework for data policies 90 6.2 Government requests for user data 96 6.3 Digital security risk management cycle 101 DN.1 Changes in Digital Adoption Index scores and per capita income, 2014–16 130 DN.2 Top Digital Adoption Index scores, 2016, and largest improvements, 2014–16 131 MAPS ES.1 Data protection and privacy legislation worldwide, 2018 xx 2.1 Internet exchange points around the world, 2018 16 B3.5.1 Night lights in India 41 5.1 Average price of 1 gigabyte of mobile data per month, by country, 2016 76 DN.1 Digital Adoption Index score, by country 124 TABLES 1.1 Data hogs: Top 10 private companies globally, by market capitalization, May 2017 2 1.2 Big data 3 2.1 Top 10 global websites 12 2.2 Internet exchange points by region 16 2.3 Average revenue per user from internet data, 2016 22 2.4 Percentage of aggregate peak period traffic by region, 2015–16 24 3.1 Examples of artificial intelligence applications for the Sustainable Development Goals 36 4.1 Typology of actors in the personal data market 54 4.2 Benefits from personal data to individual 56 4.3 Risks and remedies 60 6.1 Open data principles 92 DN.1 Data and affordability 108 DN.2 Government data infrastructure and open data 114 DN.3 Data sources for the Digital Adoption Index 125 DN.4 Digital Adoption Index and subindexes, by country and year 125 viii Contents Foreword It is my pleasure to introduce the 2018 edition in the series models, away from subscriber-funded networks to Information and Communications for Development, which advertising-funded services. This has important implica- in this fourth edition focuses on the opportunities and chal- tions for how infrastructure is financed. In this evolving lenges of data-driven development. Since 2006, Information context, we must ensure that data is used for inclusion, not and Communications for Development has been a flagship exclusion, and for enhanced privacy, not greater threats to report of a World Bank team that, this year, was elevated security. The final chapter of the report looks at data poli- to a department in its own right and changed its name from cies for the digital economy and how conflicting demands Information and Communication Technologies to Digital can be reconciled. At a time when governments around the Development. The changes reflect the World Bank’s strong world are reviewing existing data policies, and writing new push to realize the potential of digital technologies to advance ones, such as the European Union’s General Data Protection development, particularly in the poorest countries, and mark Regulation, this report seeks to contribute to the debate. a shift toward a focus on the applications of technology, As in previous years, this edition has been researched and rather than technology per se, and from supply to client drafted jointly with the World Bank’s Finance, Competitive- demand. These principles, introduced by the 2016 World ness and Innovation Global Practice. It also benefits from Development Report: Digital Dividends, guide the World Bank contributions from the International Telecommunication Group’s investment and technical assistance in this area. Union and the United Nations Global Pulse, as well as The theme of data-driven development fits well with this inputs and review from other parts of the World Bank new focus. Data is all around us, and can be invaluable once Group. We are likewise grateful for the support of the refined, processed, and analyzed. This report shows how Digital Development Partnership and its donors, who made governments in developing countries can enhance their use the report possible. of data to provide better services to citizens. It also shows Smart use of data holds immense development poten- how the business sector is starting to capitalize on data for tial that is already available to governments, businesses, competitive advantage. For citizens, the report argues that and citizens. I am confident that this report will help these new tools can allow them to take more control of personal opportunities materialize to boost economic growth, reduce data and benefit more directly from its value. For the World the digital divide, and bring better services and benefits to Bank Group and its development partners, the report the people who need them most. contains plentiful examples of how big data and open data can be harnessed for better development outcomes. Boutheina Guermazi But challenges loom. The growth of data platforms is Director, Digital Development Department changing the profile of competitive markets and business World Bank ix Acknowledgments This report was prepared by a team from the Digital • Chapter 4, Data and People: Siddhartha Raja, Tatiana Development Department (DDD) and the Finance, Nadyseva, Roku Fukui, and Rachel Firestone (DDD) and Competitiveness, and Innovation (FCI) Global Practice Michael Minges of the World Bank Group, supported by the International • Chapter 5, Data and Firms: Carlo Rossotto, Mona Telecommunication Union (ITU) and the UN Global Badran, and Tim Kelly (DDD), Elena Gasol Ramos, Eva Pulse team. The editorial team was led by Tim Kelly, under Clemente Miranda, and Prasanna Lal Das (FCI), and the guidance of Jane Treadwell (DDD), and comprised Michael Minges Siddhartha Raja and Carlo Rossotto (DDD), Prasanna Lal Das and Elena Gasol Ramos (FCI), Phillippa Biggs (ITU), • Chapter 6, Data Policies: Elena Gasol Ramos, Eva Felicia Vacarelu (UN Global Pulse), and Andrew Stott and Clemente Miranda, and Prasanna Lal Das (FCI) Michael Minges (independent consultants). The team was • Data Notes: Michael Minges and Bradley Larson (East supported by Christine Howard and Roku Fukui (DDD) Asia and Pacific Global Practice) and David Hollander (UN Global Pulse). The work was funded by the Digital Development Partnership Trust Fund, Inputs, comments, guidance, and support at various whose members include the governments of Denmark, stages of the report’s preparation were received, and two Finland, Japan, and the Republic of Korea as well as Micro- formal review meetings were held. At the project concept soft and the Global System for Mobile Communications note review meeting, held November 16, 2016, and chaired Association. Early work on the report also benefited from a by Pierre Guislain (Senior Director, Transport and Digital contribution from the World Bank Group budget. Development), the peer reviewers were Randeep Sudan The principal authors by chapter are (DDD), Holly Krambeck (Transport), and Uwe Deichmann (World Development Report 2016 Co-Director). Comments • Executive Summary: Tim Kelly and Roku Fukui were also received from Roku Fukui, Anat Lewin, Siddhartha • Chapter 1, Overview: Tim Kelly, Andrew Stott, and Raja, and Masatake Yamamichi (DDD) and Prasanna Michael Minges Lal Das and Jill Sawers (FCI). At the decision meeting of April 25, 2018, chaired by José Luis Irigoyen (Senior • Chapter 2, Data Supply: Michael Minges and Tim Kelly Director, Transport), the peer reviewers were Charles Hurpy • Chapter 3, Data for Good: Phillippa Biggs (ITU) and and Casey Torgusson (DDD), Trevor Monroe (Global Felicia Vacarelu, Miguel Luengo-Oroz, Mila Romanoff, Themes—Knowledge), and Tariq Khokhar (Development and Robert Kirkpatrick (UN Global Pulse) Indicators and Data). Additional comments were received xi from Mark Dutz, Mary Hallward-Driemeier, and Fredes- dissemination of the book. The maps, with the exception of vinda Montes Herraiz (FCI), as well as from the author Map B3.5.1 from the World Bank’s India Lights Platform, team. Inputs were also received from Juan Navas-Sabater, were drawn by Bruno Bonansea (Creative Services). The Andrew Stott, and Isabella Hayward (DDD). team would also like to thank the many other individuals, Special thanks are owed to Patricia Katayama, of the firms, and organizations that have contributed through World Bank’s Development Economics unit, and Michael their continuing support and guidance to the work of the Harrup, of the Bank’s Editorial Production team, for over- World Bank Group, particularly those focused on data for sight of the editorial production, design, printing, and development. xii Acknowledgments Abbreviations 4G fourth generation IoT Internet of Things 5G fifth generation IP internet protocol AI artificial intelligence ISP internet service provider APEC Asia-Pacific Economic Cooperation IT information technology API application programming interfaces ITU International Telecommunication Union B2B business to business IXP internet exchange point DAI Digital Adoption Index Mbps megabits per second DDD Digital Development Department OECD Organisation for Economic Co-operation and EC European Commission Development ECLAC Economic Commission for Latin America and OTT over-the-top services the Caribbean POS point of sale EU European Union SDGs Sustainable Development Goals FCI finance, competitiveness, and innovation SME small and medium enterprise FTC Federal Trade Commission (United States) UN United Nations GDPR General Data Protection Regulation UNDP United Nations Development Programme IBM International Business Machines UNDG United Nations Development Group ICT information and communication UNICEF United Nations Children’s Fund technologies UNCTAD United Nations Conference on Trade and IOM International Organization for Migration Development xiii Executive Summary Data deluge Both types of data, their potential uses, and associated risks are all growing exponentially. Figure ES.2 shows common I n a sample second in July 2018, it is estimated that sources of personal data. some 2.7 million emails were sent and received, 74,860 YouTube videos watched, and 59,879 gigabytes of inter- Who benefits? net traffic carried.1 Clearly, we generate huge and growing volumes of data. This new report, the fourth in the Information and The digital economy has become more information Communications for Development series (figure ES.3), intensive, and even traditional industries, such as oil and examines data-driven development, or how better infor- gas or financial services, are becoming data driven. By 2020, mation makes for better policies. The report aims to help forecasts Cisco (2017), global internet traffic will reach firms and governments in developing countries unlock the about 200 exabytes per month, or 127 times the volume value in the data they hold to improve service delivery and of 2005, with much of the growth coming from video and decision making and empower individuals to take more smartphones (figure ES.1). And that data may hold huge control of their personal data. The report asks just how we value. McKinsey Global Institute (2016) estimates that cross- can use this data deluge to generate better development border data flows in 2014 were worth about US$2.8 trillion, outcomes. up 45-fold in value since 2005. People’s lives can benefit greatly when decisions are The vast majority of the data that exists today was informed by relevant data that uncover hidden patterns, created in just the past few years (IBM 2013). The chal- unexpected relationships, and market trends or reveal pref- lenge is to extract value from it and to put it to work— erences. For example, tracking genes associated with certain for firms, governments, and individuals. Every citizen is types of cancer or explaining the potential links of Nean- producing vast amounts of personal data that, under derthal DNA with resistance to the common flu virus or the right protective frameworks, can be of value for the Type II diabetes can help improve treatments. As argued in public and private sectors. Firms are willing to pay ever- chapter 3, development partners therefore need to establish increasing amounts for our attention on social media sites strategies to better use data for development, while interven- and to mine the data we produce. But even data that is ing appropriately in the data ecosystem and respecting data produced unintentionally—a byproduct of other processes, protection and privacy. known as “data exhaust,” such as call data records or GPS The World Bank Group, for instance, has established coordinates—can have value when effectively analyzed. a Technology and Innovation Lab for improving data use xv Figure ES.1 The growing internet a. Internet users b. Internet traffic 4,000 100 100 250 3,578 3,385 90 90 21 19 17 3,500 3,150 27 25 23 2,880 80 80 200 3,000 2,631 11 14 16 70 70 5 7 9 2,424 2,500 2,184 60 60 150 1,991 48 2,000 43 46 50 50 40 34 37 40 40 100 1,500 29 31 30 30 1,000 20 20 50 500 10 10 68 68 68 67 67 67 0 0 0 0 10 11 12 13 14 15 16 * 15 16 17 18 19 20 17 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Internet users (millions) Managed IP (percent, left scale) Per 100 people Mobile data (percent, left scale) Fixed internet (percent, left scale) Total IP traffic (exabytes per month, right scale) Sources: ITU (panel a); Cisco (panel b). Note: An exabyte is one quintillion bytes (1 followed by 18 zeroes). Just five exabytes would be equivalent to all the words ever spoken by human beings (http://highscalability.com/blog/2012/9/11/how-big-is-a-petabyte-exabyte-zettabyte-or-a-yottabyte.html). IP = internet protocol. * Data for 2017 is an estimate. Figure ES.2 Types of personal data two editions of the Artificial Intelligence for Good, Global Summit.4 And the team from UN Global Pulse, another Health Government partner in this report, is working with UN partners to • Medical history • Identification number and identity responsibly harness big data and artificial intelligence for • Prescriptions and vaccinations • Address • Fitness tracking • Civil information (birth, marriage, development and humanitarian action (see chapter 3).5 and so on) However, firms and organizations that can make the best • Legal records use of the data are not necessarily the ones that collect it. An “open data” mind-set is critical to data-driven development. Web Mobile phone Thus, an open marketplace for data is to be encouraged • Email • Number and preferred network • Browsing and search history • Call data records within limits. It is important therefore to develop appro- • Content (social profiles, posts, • Location data (GPS) photos, and so on) • Social media contacts priate guidelines for data sharing and use, and for anonym- • Contacts, followers, friends • Purchasing history izing personal data. Governments are already beginning to release value from the huge quantities of data they hold Financial Other to enhance service delivery, though they still have far to • Accounts • Home information go to catch up with the commercial giants. To use data • Transactions • Travel • Debts • Vehicle information intelligently for better development outcomes, national • Investments • Inferred data, created using other • Insurance statistical offices will continue to play a core function, data points including that of objectivity and impartiality, producing data “without fear or favor.” But many statistics offices are in its projects, including using artificial intelligence and struggling technically and financially. To remain relevant in blockchain.2 This is part of a broader work program that an on-demand world, they need to strive for real-time data aims to leverage data and technology in its work.3 The availability, striking an informed balance between accuracy International Telecommunication Union has so far hosted and timeliness. xvi Executive Summary Figure ES.3 The Information and Communications for data—such as Google and Alibaba Global—thanks to these Development series existing infrastructures. Stock markets, in turn, assign huge potential to these data-rich companies, and undervalue the companies that keep the digital plumbing working. We have seen this pattern before. In the early part of the nineteenth century, the markets of the time afforded opti- • 2006: Global Trends and Policies mistic valuations to the companies that built railroads. But as the century drew on, railroad investors went bankrupt or were nationalized because of their huge debts, even as the companies whose products they carried, such as mail-order companies, thrived in the early twentieth century. Once again, we face an inflection point. For more than a hundred years, infrastructure companies made their money • 2009: Extending Reach and Increasing Impact primarily from subscriptions and usage charges paid by users—who paid by the minute, by the mile, and lately by the megabyte. This is changing. The value of telecom- munication networks is now not so much in data transport as in data storage. As chapter 2 shows, the companies with the highest market valuations are those that collect then monetize their customers’ data through targeted advertising. • 2012: Maximizing Mobile Services from Facebook, Google, or Tencent are largely “free” at the point of use—yet their bandwidth requirements grow ever larger, as does their customer reach. Beyond internet business or commercial applications, Note: No Information and Communications for Development report was multiple opportunities also exist for harnessing the value published in 2015, as this coincided with the World Development Report: Digital Dividends. of big data and artificial intelligence to help us achieve shared development objectives, as exemplified in chapter 3. However, global efforts to develop new frameworks for the Data-driven business models responsible use of emerging technologies must address their Companies are also developing new markets and making prof- implications for society and the consequences of both using its by analyzing data to better understand their customers. This data and algorithms, and of failing to use them. is transforming conventional business models, as explored in chapter 2. For years, users paying for calls funded telecom- Data belongs to all of us munications. Now, advertisers paying for users’ data and attention are funding the internet, social media, and other People need to exert greater control over the use of their platforms, such as apps, reversing the value flow. The share personal data. Their willingness to share data in return for of the value extracted by the network providers is shrinking, benefits (real or perceived) and free services, such as virtu- threatening future investment. Good business models for ally unrestricted use of social media platforms, varies by investment in telecommunication networks typically have country and by age group (figure ES.4). Consumer research high up-front sunk costs, but very long-term returns. Twenty from GfK, a German research institute, shows that willing- to thirty years ago, companies that built networks—such as ness to share is highest in China and lowest in Japan. Early NTT, China Mobile, AT&T, or Deutsche Telekom—were the internet adopters, who grew up with the internet and are champions of their respective national stock markets. Their now age 30–40, are the most willing to share (GfK 2017). assets, like the infrastructure that they put in place, represent Many countries and regions have taken steps recently to the backbone services operate on. But their market values have update and reinforce rules on the use of personal data. The fallen in comparison to the businesses gathering and storing European Union’s General Data Protection Regulation, Executive Summary xvii Figure ES.4 Are you willing to share your data? China 38 Mexico 30 Russian Federation 29 Italy 28 Global 27 Brazil 26 United States 25 Argentina 24 Korea, Rep. 20 Australia 17 Spain 16 United Kingdom 16 France 15 Canada 14 Percentage of positive responses to the Netherlands 12 question "Would you share personal data (financial, driving records etc.) for benefits Germany 12 (e.g., lower cost, personalization)?" Japan 8 0 5 10 15 20 25 30 35 40 Source: GfK 2017. Note: Based on more than 22,000 consumers online in 17 countries with a response of 7 (on a scale from 1 to 7), where 7 represents full agreement. which went into effect on May 25, 2018, imposes a long list Other emerging markets, such as India, Indonesia, the of requirements for companies processing personal data. Russian Federation, and Vietnam, are also seeking data local- Violations will result in fines that could total as much as ization. The Russian Federation has blocked LinkedIn from 4 percent of global annual turnover. operating in the country after the site refused to transfer data Other countries have taken steps to restrict the flow of on Russian users to local servers. Divergent rules on the treat- their citizens’ data beyond their borders (data localization). ment of data impose significant costs on doing business online. In China, where data localization is strongly championed, Business organizations, including the International Chamber restrictions on moving data are severe. Long-established of Commerce, would like to establish rules to restrain what controls over technology transfer and state surveillance of they call “digital protectionism.”6 However, a serious gap exists the population are predominant, and such measures form in global governance with regard to cross-border trade in data, part of the country’s “Made in China 2025” industrial strat- and a coherent approach is prevented by differing philoso- egy. The strategy is designed, in part, to make the country phies among the main trading blocs. a global leader in tech-intensive sectors such as artificial The ownership and control of data will continue to be a intelligence and robotics. Chinese technology giants, includ- major question for society. Broadly speaking, there are three ing Baidu, Alibaba, and Tencent, are among the biggest in possible answers to the question “Who controls our data?”: the world, and the country is establishing strong positions firms, governments, or users. No global consensus yet exists in new sectors like the Internet of Things (appliances, on the extent to which private firms that mine data about machines, and other items able to connect with the internet individuals should be free to use the data for profit and to and exchange data). Throughout the world, data is regarded improve services. Some governments argue that data from as a new asset class vital for industrial competitiveness. a country’s citizens belongs to those citizens and should xviii Executive Summary Figure ES.5 Toward a new value chain for personal data Government, business, and Data protectors organizations Volunteered Individuals Data collectors Data brokers Data users Produce Observed Collect and monetize Aggregate, analyze, and monetize Add value and monetize Unpaid and paid digital services Advertising Direct and indirect benefits not leave the country without permission. Data dependency monopoly profits, unless competition rules are modified leads to new risks of exclusion. The data poor, who leave to deal with new concepts of dominance. The emergence little or no digital trail because they have limited internet of multisided platforms, explored in chapter 5, poses new access, are most at risk of exclusion. But, equally, those who challenges for regulators. live in ways that society deems unconventional may also risk Data and the internet have predominantly been regarded by exclusion, for instance, because they lack a digital ID or are pioneers and campaigners as a decentralized, self-regulating considered an insurance risk. community. Activists have tended to regard government This report espouses the view that citizens should control intervention with suspicion, except for its role in protecting their own data and should be free to choose how to release personal data, and many are wary of legislation to enable data it and even to commercialize it (figure ES.5), as explored in flows. But that position is under pressure from the increasing chapter 4. centralization of the internet (Economist 2018) and a series The growth of the data economy therefore requires of revenue data breaches and media exposés of questionable changes in competition policy and the regulation of business practices by social media platforms. The use by privacy. In a traditional, or one-sided, market, dominant political parties in Kenya, the United States, and elsewhere of firms are bad for overall market development. But when data harvested from social media profiles does not appear to it comes to personal data, splitting the market share have broken any rules, but it has led politicians on both sides too many ways may inconvenience users and compli- of the Atlantic to take a closer look at social media giants, cate matters for the individual if the different platforms such as Facebook and Twitter.7 The proliferation of “fake do not connect, or if they require different passwords. news” has also spurred calls for action.8 As data becomes more important in shaping markets, Data collected by governments, and thus paid for by it may reinforce tendencies toward monopoly, and thus taxpayers, arguably belongs to all of us. But there are limits Executive Summary xix Map ES.1 Data protection and privacy legislation worldwide, 2018 Legislation Draft legislation No legislation No data IBRD 43802 | SEPTEMBER 2018 Source: UNCTAD (http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom-Data-Protection-Laws.aspx). to the openness paradigm. Citizens may not want data about these principles to firms from other parts of the world that themselves to be exposed without protection. And govern- wish to do business in Europe. ments often lack the resources to extract value from their Ironically, although data is becoming ever more impor- data without private partners. Data-driven development tant, data about data is still hard to find. The Data Notes needs greater dialogue between the custodians of a country’s to this report set out some of the indicators that should data and its users. The key to unleashing the power of data- exist and present data that do exist on an internationally driven development for developing countries lies in intelli- comparable basis for indicators such as the price and gent management, use, and supervision of data. affordability of data transmission and the availability of Chapter 6 reviews data-related policy issues relevant to open government data. the digital economy. It considers policies geared toward This report aims to stimulate wider debate within the building consumer trust, policies that facilitate or can affect development community on the nature of data for develop- access to data, and the use of data as infrastructure. The ment. It is not the first word on this topic and certainly will chapter also covers mainstreaming policies, such as those not be the last. But it is a topic of growing importance that that facilitate the use of data for innovation or those that cannot be ignored. build digital skills. At least 35 economies are currently draft- ing data protection laws (map ES.1). In addition, a number of economies are considering reforms to their legal frame- Notes works. One factor driving this consideration is the European 1. Internet Live Stats (Internetlivestats.com), one second of Union’s adoption of the General Data Protection Regulation. traffic on July 31, 2018. While the regulation introduces, or confirms, many import- 2. Blockchain is technology that serves as a decentralized ant principles for data protection and privacy, it also extends digital ledger that provides immutable record keeping. xx Executive Summary Applications are emerging in land registries, money remit- /solutions/collateral/service-provider/visual-networking tances, biometric ID, and so on. -index-vni/complete-white-paper-c11-481360.pdf. 3. See http://blogs.worldbank.org/taxonomy/term/15718 for Economist, The. 2018. “How to Fix What Has Gone Wrong with information. the Internet.” Special report, June 28. https://www.economist 4. See https://www.itu.int/en/ITU-T/AI/2018/Pages/default.aspx. .com/special-report/2018/06/28/how-to-fix-what-has-gone -wrong-with-the-internet. 5. For information see https://www.unglobalpulse.org/pulse-labs. GfK. 2017. “Willingness to Share Personal Data in Exchange 6. See https://iccwbo.org/publication/trade-in-the-digital-economy/. for Benefits or Rewards.” https://www.gfk.com/fileadmin 7. See https://www.theguardian.com/news/series/cambridge /user_upload/country_one_pager/NL/images/Global-GfK -analytica-files. _onderzoek_-_delen_van_persoonlijke_data.pdf. 8. See https://www.digitaltrends.com/mobile/google-news International Business Machines Corporation (IBM). 2013. -initiative/. Harness the Power of Big Data: The IBM Big Data Platform. New York: McGraw-Hill. McKinsey Global Institute. 2016. Digital Globalization: The New References Era of Global Flows. New York. https://www.mckinsey.com Cisco. 2017. Cisco Visual Networking Index: Forecast and /business-functions/digital-mckinsey/our-insights/digital Methodology, 2016–2021. https://www.cisco.com/c/en/us -globalization-the-new-era-of-global-flows. Executive Summary xxi Chapter 1 Data: The Fuel of the Future Data, data, everywhere With some justification, therefore, data has been called the new gold (EC 2011), the new oil (Toonders 2014), or the A self-driving car, one of the most anticipated world’s most valuable resource (The Economist 2017). Like developments of the next decade, is expected to oil, unprocessed data has relatively little value and needs to be generate some 4,000 gigabytes of data for each mined, refined, stored, and sold on to create value—albeit in hour of driving, according to chip maker Intel (Nelson 2016). data centers rather than in oil rigs. But unlike oil, the quantity To put it another way, just 3 million autonomous vehicles of data is ever increasing, not diminishing. Even though data would generate, or consume, more data than the combined is a nonrivalrous good, in the sense that my consumption of human population of more than 7 billion. Vehicles provide it does not affect yours, it is also excludable, which means it just one example of how data generation and use are growing can be sold for profit, many times over. This makes it what explosively. Other machines generating an overload of data economists sometimes call a club good, like privately owned include satellites, environmental sensors, security cameras, safari parks or pay-per-view television. But because of the and, of course, the ubiquitous mobile phone. ever-increasing quantity of data, extracting value from it We are undoubtedly experiencing a data revolution (UN requires ever-greater computer power. Thus, the spoils from Data Revolution Group 2014) in which our ability to gener- data-driven markets typically go to the largest players; those ate, process, and utilize information has been magnified with the deepest pockets, the most users, the largest data many times over by the machines that we increasingly rely centers, and most wide-ranging ability to collect and analyze upon. By 2016, according to IBM, some 90 percent of data data. Consequently, it is possible for market capitalization that exists had been created within the previous 12 months, in companies like Facebook, Tencent, or Alibaba to exceed a rate of 2.5 quintillion bytes per day (IBM 2016). Firms are their annual revenue by 15 times or more and for the market increasingly finding hidden value in some of that data. Some capitalization of Apple and Amazon to touch the US$1 tril- 7 of the top 10 companies worldwide, by market capitaliza- lion mark in mid-2018, because investors view them as well tion, are data driven in that they create value primarily from positioned to take advantage of future data trends. the data they collect from or sell to their customers. The remaining 3 firms in the top 10—in the more traditional How data is changing development financial services, energy, and health care sectors—also increasingly build data into their products and services or This report is about how the data revolution is changing the use it to improve them (table 1.1). behavior of governments, individuals, and firms. Specifically, 1 Table 1.1 Data hogs: Top 10 private companies globally, by market capitalization, May 2017 Market capitalization 2016 revenue Rank Company Country (US$ billions) (US$ billions) 1 Apple United States 801 218 2 Google / Alphabet United States 680 90 3 Microsoft United States 540 86 4 Amazon United States 476 136 5 Facebook United States 441 28 6 Berkshire Hathaway United States 409 215 7 Exxon Mobil United States 346 198 8 Johnson & Johnson United States 342 76 9 Tencent China 335 22 10 Alibaba China 314 21 Top 10 total 4,684 1,090 Data-driven companies as percent of top 10 76.6 55.1 Source: Adapted from Meeker 2017, with market capitalization data from CapIQ and valid for May 26, 2017. Note: Data-driven companies are shown in red. the report examines how these changes affect the nature A data typology of development—economic, social, and cultural. How can Data-driven development is an emerging and rapidly governments extract value from data to improve service deliv- changing field. So it may be useful, at the outset, to define ery in the same way that private companies have learned to terms recurring across the chapters. These are not fixed do for profit? Is it feasible for individuals to take ownership of or official definitions, but rather working usage for this their own data and to use it to improve livelihoods and quality report: of life? Can developing-country firms compete with the inter- net majors on their own turf and even be more innovative • Big data, a commonly used term, describes data sets so in their use of data to serve local customers better? Several large or so complex that traditional data processing tech- potential audiences could therefore benefit from this report: niques are inadequate. The field of big data analytics uses advanced computational techniques to extract mean- • The primary audience is government policy makers, though ingful information (such as patterns, trends, repeti- not in a single line ministry, such as information and tion) from data. For the moment, big data is largely communication technology or finance, but rather across the domain of large private companies. But as tools to government, given that data is a multidisciplinary concern. mine it become cheaper and more readily available, • A secondary audience would be individuals concerned about smaller companies and governments will also use big how their personal data is used and those interested in how data. It can be useful to further distinguish between big the data revolution might impact future job prospects. data produced intentionally or unintentionally and that • Beyond that, private sector firms, particularly in devel- produced by humans and by machines, as in table 1.2. oping countries, looking to expand their markets and Data exhaust, which is unintentionally created by improve their competitive edge will find interesting º humans. This can include metadata (data about data), examples of how other firms are doing that. such as call-data records derived from mobile phones, • Finally, development professionals should find the report or the trail of data left by users engaged in other activ- relevant as they seek to use data more creatively to tackle ities, such as keystrokes. Data exhaust generally has long-standing global challenges, such as eliminating low value, but the trail left by millions of users can be extreme poverty, promoting shared prosperity, or miti- mined or combined to extract value or to hack into an gating the effects of climate change. otherwise secure system. 2 Information and Communications for Development 2018 Table 1.2 Big data increase the risk and uncertainty about how private data can be used in the future (Tucker 2017). Data generation Intentional Unintentional • Open data refers to data made freely available and Data Human Primary content Data exhaust agent deliberately stored in an easily read data format, particu- Machine Secondary content Internet of Things data larly by other computers, and thereby repurposed. For instance, data on airline schedules could be easily º Internet of Things data, which is intentionally cre- read by travel companies to generate customized itin- ated, but from sensors and other internet-connected eraries for travel websites. Governments may use open devices, rather than from humans. This mainly has data to promote transparency and accountability in their value in the aggregate—and over time—but can also operations and allow voters to measure the performance be used to provide alerts for impending events, such of different government functions. As a philosophy, as extreme weather conditions. therefore, open data is intended to encourage the juxta- Primary content is intentionally created by humans, typ- position of data from different sources to create value and º new applications. It is estimated, for instance, that some ically users. An example here might be a social media profile or a browser search history. When thousands or 500 different applications use London transport data more of these are combined and anonymized they can (World Bank 2014), and the savings to the UK economy be used, for instance, for analyzing popular or emerging from its open data policy amount to some £6.8 billion trends. Humans also create primary content in the form (about US$9.5 billion) a year (Government of the United of videos, academic papers, blogs, and the like that can Kingdom 2013). Open data tools are particularly useful be mined, for instance, for sentiment analysis. in the transport sector (see box 1.1). Secondary content is intentionally created, but through • Metadata, or “data about data,” is used to classify, catego- º artificial intelligence rather than directly by humans. rize, and retrieve data files. For instance, metadata might A benign example would be a chatbot that helps a user include the date on which it was created, the number fill in a form online by giving suggestions. A malign of pages or data size, and keywords that can be used to example would be a fake social media profile that search. Data attributes may be added to data according to seeks to influence buying habits or political opinion. the way it is typically used, for instance, how popular it is as a function of how frequently it is downloaded. Meta- • Personal data relates to an individual and is generally data helps with data analysis and can be applied to data concerned with private information. Personal data can users, such as by giving them attributes, sometimes based form large, complex data sets (such as multiple health on inferred data, that equate to a “reputation.” indicators including weight, blood pressure, or heart rate measured over a lifetime) but more normally constitute • Data platforms offer a convenient and cost-effective small data, which can be easily monitored. Personal data way to link customers and suppliers. Some platforms may be willingly exchanged in return for convenience may connect only peers (such as a dating website) and (such as a phone number or email address), but it can others might be internal to an organization (such as an also be given away unwittingly (such as date of birth intranet). But most of the popular platforms using the provided to enter an online competition) or unwillingly internet are multisided platforms. Uber, for instance, (such as data hacked from a personal email account). connects drivers with riders; AirBnB connects property The consequences of loss of personal data, explored more owners with guests; and Jumia connects sellers with in chapter 4, might include loss of privacy and loss of potential buyers. But the biggest platforms are those that control (over the future use of personal data) and a loss connect advertisers with consumers, usually in return for of agency (such as being exposed to a more limited range some kind of free service, such as social media or web of news sources or opinions as a result of previously browsing. As explored in chapter 5, multisided platforms, expressed preferences). What is relatively new is how driven by advertising, are now among the most powerful persistence, repurposing, and spillovers from big data firms in the world. Data: The Fuel of the Future 3 Box 1.1 Open data tools for improving transport through big data Rise in digital data. Digital data has proliferated with the rapid increase in smartphone owner- ship in advanced and emerging market economies, alongside advances in global positioning systems and digital sensors. This data has the potential to transform transport systems world- wide. The location-tracking data provided by smartphones, for instance, can reveal how and why people travel, information critical for optimizing transport networks. Accordingly, open- source tools and cloud-based platforms have been developed to help collect, manage, and analyze the ever-increasing volume of digital data. These easily accessible tools provide indi- viduals, governments, and private entities with sophisticated analysis capabilities, empowering them to improve all aspects of transport. Such tools will be particularly beneficial in developing countries that have limited resources. Open-source tools. The World Bank has developed a variety of free-of-charge tools that capitalize on big data to facilitate transport-related development projects across the globe (see Figure B1.1.1). These tools provide numerous capabilities, including transit system analysis, route planning, and road condition and incident reporting. Open Transit Indicatorsa allows public transit administrators to assess existing services and identify improvements through the collection and analysis of standardized transit data. This approach has been used to address transit problems in China, Kenya, Mexico, and Vietnam, among others.b The Rural Accessibility Platform uses freely available OpenStreetMap data to evaluate the accessibility of rural population centers to points of interest.c Indices of rural accessibility have been used to identify needed transportation improve- ments in countries including Bangladesh, the Lao People’s Democratic Republic, and Zambia.d These open-source data and tools make transport analysis accessible for a broad range of users. Citizen engagement. The increasing ubiquity of smartphones and internet connectivity is allowing individuals to provide valuable data and contribute to development efforts. Citizen engagement is prioritized in many of the World Bank’s transport-focused open-source tools. For example, the smartphone application RoadLab uses a crowdsourcing approach to obtain route information and roadway infrastructure conditions from users (Wang and Guo 2016). Figure B1.1.1 How open data tools can assist transport Cloud server High-tech, Smart phone low-cost, Open-source GIS platform easily accessible ITS tools Web access Open-source ITS tools ITS analysis Crowd sourcing Citizen engagement Government accountability Note: GIS = geographic information system; ITS = intelligent transport systems. (continued next page) 4 Information and Communications for Development 2018 Box 1.1 (continued) The related tool RoadLab Pro was used to assess the conditions of unclassified road networks in Mozambique (Espinet, Wang, and Mehndiratta 2017), demonstrating the potential of citizen- provided smartphone data in transport planning. These tools provide an easy-to-implement way for traffic engineers to obtain roadway information, particularly when professional pavement testing equipment and base geographic information system maps are not available. Similarly, DRIVER capitalizes on crowdsourced data to collect road incident information, which can then be visualized and analyzed to improve enforcement and resource allocation. In the Philippines, DRIVER has been applied to identify and prioritize problematic road areas for interventions (World Bank 2016). Sources: Progress Analytics LLC; Juan Navas-Sabater. a. For more information, see https://www.transitwiki.org/TransitWiki/index.php/Open_Transit_Indicators. b. http://www.digitalmatatus.com/about.html; http://www.worldbank.org/en/news/feature/2013/11/05/mexico-city-open -database-improves-transit-efficiency-helps-commuters; Krambeck and Qu 2015. c. https://github.com/WorldBank-Transport/ram-backend. d. https://developmentseed.org/projects/ram/; Iimi et al. 2016. How governments use data different channels, and among different government institutions From e-government to digital government How governments use data runs throughout this report, • Allowing governments to become more “data driven” at all although because it is the subject of a separate World Bank levels, from policy making through operational manage- Group report (World Bank 2017) it gets no separate chapter ment and risk management to individual decision making here. Chapter 3 nonetheless focuses on big data for social good, by international, nongovernmental, and humanitarian • Underpinning the creation of “smart cities” (and “smart organizations, as well as by governments. nations”) whose systems and infrastructure adapt auto- In the first generations of “e-government,” much of the matically to the needs and behaviors of their inhabitants emphasis was on channels—using web browsers, and more • Providing, through “open data” and other programs, recently, mobile devices, to access government information authoritative reference, geospatial, and other data to the and services and to perform transactions. In this period, national economy and society as a whole to improve data was often seen as just the payload of the transaction— transparency, to enable economic growth and business information supplied by the citizen or the business to innovation, and to increase the engagement of citizens support the request for service and the information supplied in the co-creation of public services. An example of by the government in return. this would be a National Spatial Data Infrastructure, or However as “e-government” has evolved into “digital Digital Maps. government,” data is seen increasingly as a strategic asset with value lasting beyond a particular transaction and able Viewing government data as a strategic asset leads to to strategically transform the efficiency and effectiveness of requirements for effective and strategic data governance and government through: data management across the entire life cycle of data, includ- ing how data is collected, described, and catalogued, as well • Making “e-government” transactions more attractive and as secured and controlled (not just to protect confidential- useful to citizens and businesses by eliminating the need ity, but also to ensure availability and integrity). Preparing to supply the same information again and again, making government data for wider use will require elimination of transactions more suitable for the mobile channel, and unnecessary duplication or avoidance of re-collection of allowing continuity of transactions over time, through data. It will also require a strategic view of how data is shared Data: The Fuel of the Future 5 across government, used within government and other public too often neglected in government, especially at the regional services, and made available to other economic and societal or local level. actors to generate additional economic and social value. Sharing data across government The changing role of national statistical offices Data collected and held by one government agency may be These requirements are in turn leading to demands for valuable to another agency in its operations. For instance, it new skills and roles, including “data scientists” and “chief may relieve the second agency of responsibility for collecting data officers,” and new functional capabilities such as the data itself. And in countries such as Belgium, Estonia, “data analysis,” “big data,” and “visualization.” Historically, or the Russian Federation the government is not allowed to national statistical offices were, appropriately, the central ask citizens again for data that it has already collected from repository of data, along with national archives, as they them. have the skills and resources to catalogue and manage data. Of course, if personal or classified data is shared between The skills of information scientists and librarians in these ministries, it is important that it is shared securely. In the offices may not be so readily available to more casual users United Kingdom, two compact discs containing personal of data in line ministries. But national statistical offices details of some 25 million children were lost in transit have had to reinvent themselves in the internet age, in between two government agencies (Government of the which a simple web search can come up with many more United Kingdom 2016). This led to the mandatory use of possible sources of information than even the most dedi- encryption when moving confidential data between govern- cated librarian can track. ment agencies. International authorities also need to collaborate on A number of countries have taken the concept of data standards for information management. The General Statis- sharing further by explicitly recognizing the importance of tical Business Process Model, for example, is a framework unified databases accessible to and used across the public for organizing business processes in statistical organizations sector, rather than each agency keeping its own records. In adopted in more than 60 countries.1 But as national statis- 2012, Denmark published a strategy for “Good Basic Data tical offices modernize and partner with nongovernmental for Everyone—A Driver for Growth and Efficiency.”2 Public entities that provide data for official statistics, there is recog- authorities in Denmark register various core information nition that such frameworks and business models may be about individuals, businesses, real properties, buildings, too rigid. For instance, Statistics New Zealand’s 2020 strat- addresses, and more. This information, called basic data, egy affirms the organization’s role as a producer of official is seen as important for reuse throughout the public statistics but moves beyond this to acknowledge its place in sector because it is an important basis for public author- a broader ecosystem and its renewed purpose of “adding ities to perform their tasks properly and efficiently, “not value to New Zealand’s most important data” through least because an ever-greater number of tasks have to be increased data cooperation, integration, and analysis. performed digitally and across units, administrations and The challenge for national statistical offices, therefore, is sectors.” Some of the registers do not contain personal to ensure the information they hold is properly catalogued information and are released as open data (for instance, (metadata) and easily searchable, and to offer this expertise addresses). In the Netherlands. there is a similar initiative throughout government. This requires a changing business for the sharing of 17 “base registers.” The United King- model for the offices. They can no longer expect to cover dom, despite past political controversy, is collaboratively costs primarily through sale of publications, though this developing a data-sharing policy that will allow the use of may still provide an additional source of income. Instead, key databases across the public sector and, in some circum- they must rely on the central treasury for most of their stances, beyond (Government of the United Kingdom, n.d.). income, and on payment for services provided to other parts In federated countries those data sets need to be available of government. Where the central treasury is already over- not just between national agencies, but also regional and stretched, as in many developing countries, national statis- municipal agencies. Since changes to master data may first tical offices frequently struggle. Thus, just as the value of data be notified to other agencies, robust processes are essential is becoming all the more evident in the private sector, it is for the maintenance of the master data using notifications 6 Information and Communications for Development 2018 of change as early as possible; this is even more important Part III of the report brings together the policy implications in federated systems, where important changes, such as of for developing country stakeholders: address, may well be notified locally first. • Chapter 6 discusses the policy issues surrounding the use of This also exemplifies the increasing extent to which data, notably over privacy, data localization, and security leading governments see databases, not functions, as the key issues. The chapter also considers the value of digital asset of government administration, along with developing ID systems, which many countries have adopted in strategic plans to introduce interoperability standards and recent years, though some have specifically rejected them. middleware that allow seamless integration of these data- Finally, the chapter returns to the themes of open data bases through open application programming interfaces and big data and offers recommendations. (widely known as APIs). The Data Notes appendix to the report looks at statistical indicators associated with the use of data. It also presents the Structure of the report 2018 update of the Digital Adoption Index (DAI), a compos- Following this overview chapter, with its focus on govern- ite indicator first introduced in the 2016 World Development ment use of data and presentation of definitions, Part I of Report: Digital Dividends. The DAI is an analytical tool that the report looks at the “supply side” of the data sector. compares the relative adoption of digital technologies by governments, people, and firms within a country. • Chapter 2 looks at data connectivity and capacity, consid- ering where data comes from, how it is stored, and where Notes it goes. Specifically, the chapter looks at the technological drivers that make data ever cheaper to collect, store, 1. See https://ec.europa.eu/eurostat/cros/content/gsbpm-generic and transmit, and the relationship between data and -statistical-business-process-model-theme_en. economic growth. 2. To see the strategy, go to http://www.digst.dk/Home /Servicemenu/English/Digitisation/Basic%20Data. • Chapter 3 examines data technology, specifically big data analytics and artificial intelligence, and how this is References contributing to development, especially in humanitarian The Economist. 2017. “The World’s Most Valuable Resource Is interventions. The enthusiasm for the uses of these new No Longer Oil, but Data.” May 6. https://www.economist tools is tempered by awareness of the ethical issues. .com/news/leaders/21721656-data-economy-demands-new -approach-antitrust-rules-worlds-most-valuable-resource. Part II looks at the “demand side” of the data sector: Espinet, X., W. Wang, and S. Mehndiratta. 2017. “Low-Budget • Chapter 4 looks at people’s use of data and asks whether Techniques for Road Network Mapping and Road Condition scope exists for a new model for a data market in which Assessment That Are Accessible to Transport Agencies in individuals may be able to trade access to their personal Developing Countries.” Transportation Research Record: Journal of the Transportation Research Board no. 2634: 1–7. data. The underlying principle is that the data itself has no value, but the use of it has. The chapter goes on to European Commission (EC). 2011. “Data Is the New Gold.” Opening Remarks, Press Conference on Open Data Strategy, examine the potential costs of a data market in possible December 12. https://ec.europa.eu/digital-single-market/en losses of privacy, control, and agency. /news/data-new-gold. • Chapter 5 examines how firms use digital platforms in Government of the United Kingdom. 2013. “Shakespeare Review: the data economy, and how that contributes to competi- An Independent Review of Public-Sector Information.” May 15. https://www.gov.uk/government/publications/shakespeare tiveness, particularly for small and medium enterprises -review-of-public-sector-information. (SMEs). The chapter details several developing-country ———. 2016. “Review of Information Security at HM Revenue platforms and emerging business models, and concludes and Customs.” http://webarchive.nationalarchives.gov by considering how SMEs in developing countries can .uk/20100407163917/http://www.hm-treasury.gov.uk/d make better use of data to improve competitiveness and /poynter_review250608.pdf. thereby compete against the dominant international ———. n.d. “Data Sharing in Government.” Blog. http://datasharing social media companies. .org.uk/index.html. Data: The Fuel of the Future 7 Iimi, Atsushi, A. K. Farhad Ahmed, Edward Charles Anderson, Tucker, C. 2017. “Privacy, Algorithms and Artificial Intelligence.” Adam Stone Diehl, Laban Mayo, Tatiana Peralta Quiros, and Working Paper 14011, National Bureau of Economic Kulwinder Singh Rao. 2016. “New Rural Access Index: Main Research, Cambridge, MA. http://www.nber.org/chapters Determinants and Correlation to Poverty.” Policy Research /c14011.pdf. Working Paper 7876, World Bank, Washington, DC. UN Data Revolution Group. 2014.“A World That Counts: Mobilising International Business Machines Corporations (IBM). 2016. the Data Revolution for Sustainable Development.” http:// “IBM10 Key Marketing Trends for 2017 and Ideas for Exceeding www.undatarevolution.org/wp-content/uploads/2014/12/A Customer Expectations.” https://www-01.ibm.com/common -World-That-Counts2.pdf. /ssi/cgi-bin/ssialias?htmlfid=WRL12345USEN. Wang, W., and F. Guo. 2016. “RoadLab: Revamping Road Krambeck, H., and L. Qu. 2015. “Toward an Open Transit Service Condition and Road Safety Monitoring by Crowdsourcing Data Standard in Developing Asian Countries.” Transportation with Smartphone App.” Paper presented at the 95th Annual Research Record: Journal of the Transportation Research Board Meeting of the Transportation Research Board, Washington, no. 2538: 30–36. DC, January 10–14. Meeker, Mary. 2017. Internet Trends 2017: Code Conference. World Bank. 2014. “Open Data for Economic Growth.” Transport Annual Report. Menlo Park: Kleiner, Perkins, Caufield, Byers and ICT Global Practice, Washington, DC. http://docu (KPCB). https://www.kleinerperkins.com/perspectives ments.worldbank.org/curated/en/131621468154792082 / internet-trends-report-2017. /pdf/896060REVISED000for0Economic0Growth.pdf. Nelson, P. 2016. “Just One Autonomous Car Will Use 4,000 GB ———. 2016. “Philippines: Real-Time Data Can Improve Traffic of Data/Day.” Networkworld, December 7. https://www Management in Major Cities.” Press Release, April 5. http:// .networkworld.com/article/3147892/internet/one-autonomous www.worldbank.org /en/news/press-release/2016/04/05 -car-will-use-4000-gb-of-dataday.html. / philippines-real-time-data -can-improve-traffic-manage Statistics New Zealand. 2016. “Statistics New Zealand’s Strategic ment-in. Intentions for the Period 2016/17–19/20 Annual Report for ———. 2017. “Big Data in Action for Government: Big Data the Year Ended 30 June 2016.” Wellington. https://www.stats Innovation in Public Services, Policy, and Engagement.” .govt.nz/about-us/corporate-publications/. Solutions Brief. World Bank, Washington, DC. http:// Toonders, J. 2014. “Data Is the New Oil of the Digital Economy.” documents.worldbank.org/curated/en/176511491287380986 Wired. https://www.wired.com/insights/2014/07/data-new /Big-data-in-action-for-government-big-data-innovation-in -oil-digital-economy/. -public-services-policy-and-engagement. 8 Information and Communications for Development 2018 Chapter 2 Supply: Data Connectivity and Capacity The ever-expanding data universe number of websites providing partial information about T he rapid growth of internet users and faster content growth on the internet. According to Netcraft, a lead- network speeds is driving an avalanche of elec- ing research firm covering the internet, 170 million websites tronic data. About 3.5 billion people globally were were active in June 2016, up from just 8 million in June 2000 using the internet in 2017, up 73 percent, or 1.5 billion, since (Netcraft 2017). Worldwidewebsize.com puts the number 2010 (figure 2.1, panel a), and penetration has risen to almost of indexed web pages at 4.5 billion.2 Although useful, these half the world (48 percent in 2017). The rapid increase in numbers still lack the ability to portray the full scale of data users is driving demand for more internet content. accessible over telecommunication networks. They do not End-user internet speeds are also increasing rapidly, in include the so-called dark web, ranging from innocuous turn driving use of broadband content and applications private sites collecting sensor data to nefarious sites carrying (figure 2.1, panel b). Global average wired broadband out illegal or semi-legal activities. Furthermore, not all data speeds are projected to nearly double from 25 megabits going over the internet is from websites; it can also arise out of per second (Mbps) in 2015 to 48 Mbps by 2020 as more voice over internet protocol, video conferencing, gaming, and users move to fiber and higher-speed coaxial cable. As machine-to-machine communication. speeds rise, so does demand for video content.1 Mobile More is known about data flows over the internet. In broadband speeds, which are much lower than fixed speeds, the past, separate networks existed for specific content and averaged just 2 Mbps in 2015. This will more than triple functions: for instance, telecommunications for voice and, to 6.5 Mbps by 2020 as more users switch to fourth- later, text messages; broadcasting for television and radio; generation technologies. Mobile speeds vary greatly by and private networks for businesses. The development of device; smartphones are nearly three times faster than the the internet and internet protocol (IP) communications has global average, which results in more time spent online. changed all that. Communications networks have generally In the United Kingdom, time spent on the internet more shifted from circuit-switched to packet-switched IP networks, than doubled between 2005 and 2015 from 10 hours a week enabling virtually any type of content, from voice to text to in 2005 to 23 hours in 2015 (Ofcom 2016a). multimedia, to be encoded and distributed digitally. Accord- Data can either be measured as stock (the amount of data ing to information technology (IT) company Cisco, traffic stored in a location) or flow (amount of data transmitted over the internet will grow by more than 20 percent a year from one location to another). One stock indicator is the between 2015 and 2020. This data deluge has popularized 9 Figure 2.1 Internet users and broadband speeds a. Global number of individuals using the internet b. Average global broadband speeds 4,000 100 60 3,578 3,385 90 3,500 3,150 50 47.7 80 2,880 43.5 3,000 2,631 70 38.5 Megabits per second 2,424 40 2,500 2,184 60 32.8 1,991 29.5 2,000 48 50 30 46 24.7 43 40 40 1,500 37 34 20 31 30 29 1,000 20 10 6.5 500 3.9 5.1 10 2.0 2.4 3.1 0 0 0 2010 2011 2012 2013 2014 2015 2016 2017 2015 2016 2017 2018 2019 2020 Internet users (millions, left scale) Per 100 people (right scale) Fixed Mobile Sources: Cisco; International Telecommunication Union. Note: 2017 data is an estimate. a new vocabulary of petabyte and exabyte that spell checkers by more than 30 percent a year so that, by the year 2020, have not yet caught up with. Cisco proclaimed that the world it will account for 82 percent of consumer traffic and 57 entered the zettabyte era (an amount equivalent to 250 billion percent of total IP traffic. Online gaming is projected to be DVDs) in 2016 when annual global internet traffic surpassed the fastest-growing traffic stream between 2015 and 2020, 1 zettabyte (Barnett 2016). at 47 percent per year. However, it accounts for a tiny share It is useful to understand how internet traffic is classified of total consumer traffic and its contribution will only rise to understand how devices, users, applications, and services from 0.2 percent to 0.4 percent. are driving this growth. Internet traffic consists of IP and managed IP traffic. The former is exchanged between inter- Goodbye data carriers, hello data net service providers (ISPs), whereas the latter is end to end creators within the same ISP’s network. IP traffic can be further disaggregated by whether it emanates from fixed or mobile The rise of the internet has altered communications network networks.3 The two accounted for three-quarters of internet value chains. In the past, little value was perceived in the traffic in 2016, with fixed making up more than 90 percent content of traffic carried over communications networks. In (figure 2.2, panel a). Managed IP traffic is forecast to decline the telecommunication world, this traffic was mainly tele- by 10 percentage points between 2015 and 2020, and the phone calls. While some of the calls may have triggered wealth, share of mobile data in total traffic is projected to rise from the direct income accrued to telecommunication carriers that 5 percent in 2015 to 16 percent by 2020. transmitted the calls and billed them. In the case of broadcast Businesses and consumers generate traffic, with the and private networks, content was more financially significant latter accounting for more than 80 percent in 2015, a share but intra-industry (such as broadcast transmissions to satellite not projected to change much through 2020. Video domi- and cable television companies, banking transactions). nates consumer IP traffic. It accounted for 38 exabytes a The development of the internet in the 1960s, the World month of traffic in 2016, 71 percent of consumer IP traffic, Wide Web in the 1990s, and its iteration in Web 2.0 more and 43 percent of total IP traffic. It is forecast to grow recently has modified the way content is obtained and created. 10 Information and Communications for Development 2018 Figure 2.2 Global IP traffic and global consumer IP traffic a. Global IP traffic b. Global consumer IP traffic (percent) 100 250 100 90 21 19 17 90 27 25 23 80 200 80 9 11 14 16 70 5 7 70 60 150 60 50 50 40 100 40 81 82 75 78 68 71 30 30 20 50 20 10 10 68 68 68 67 67 67 0 0 0 2015 2016 2017 2018 2019 2020 2015 2016 2017 2018 2019 2020 Managed IP (percent, left scale) Mobile data (percent, left scale) Internet video Web, email, and data File sharing Fixed internet (percent, left scale) Total IP traffic (exabytes per month, right scale) Source: Adapted from Cisco Visual Networking Index Global IP Traffic Forecast. Note: IP = internet protocol. Traditional content providers such as the media and audio- Many of the world’s top websites (ranked by a combi- visual companies have moved online either with their own nation of users and page views) are platforms for user- websites or by licensing content to streaming platforms. Take generated content such as social network posts, video the BBC, which has 98 million global internet users viewing sharing, blogs, and collaboration (for example, Wikipedia) 1.5 billion pages a month.4 Or O Globo, one of Brazil’s largest (table 2.1). All of the top sites are headquartered in either newspapers, whose online readers (23 million) outnumber the United States or China, the two countries with the most print readers (300,000) by 75 times.5 A big difference is that internet users (more than 900 million combined, or just over not only can anyone access content anywhere on the public a quarter of all internet users).7 While the US sites are mostly internet, they can also create it. Users become creators by shar- global, Chinese ones are mainly local. The concentration of ing their own content with others through blogs, videos, social so much data on so few sites is concerning, particularly as networking posts, and product and service reviews. Attention the giant internet companies behind most of them branch has shifted from the carrier of the data to the creator; from into other domains. Many aspire to be the single window to “the medium is the message”6 to the messages delivered over communications, news, and shopping. the medium. A company’s telephone number is arguably no longer as important as its website, and individuals increasingly Data centers: Greener and further away, or closer to exchange their email or social networking links. Similarly, in home? video entertainment, power is shifting from the company The growth of internet content is driving the need for broadcasting the content to the creator. This is reflected in places to store it. A data center is a location with networked the rise of companies offering internet-delivered video such computers providing remote storage, processing, and distri- as Amazon and Netflix and television content creators now bution of data. They are mainly operated by global IT embracing the internet (HBO NOW streaming service). companies, governments, and enterprises that host other Supply: Data Connectivity and Capacity 11 Table 2.1 Top 10 global websites Daily time on Daily page Percent of site (minutes: views per traffic from Total sites Users Site Description seconds) visitor search linking in (millions), 2016 1 Google Internet portal 8:45 8.63 2.3 3,011,003 ~1,000 2 YouTube Video sharing 9:23 5.40 8.6 2,347,245 ~1,000 3 Facebook Social network 13:56 5.32 4.4 7,278,321 1,860 4 Baidu Search engine 7:43 6.68 4.5 118,000 657 (2015) 5 Wikipedia Encyclopedia 4:26 3.31 36.9 1,287,362 374 (2015) 6 Yahoo Internet portal 4:28 3.90 5.3 529,800 650a 7 Qq Instant messaging 5:05 4.52 3.7 211,248 877 8 Taobao E-commerce 8:33 4.48 3.8 48,973 407 (2015)b 9 Reddit News links 13:31 9.28 12.3 416,267 234 10 Tmall E-commerce 5:51 3.45 1.0 8,642 407 (2015)b Source: Adapted from Alexa and Company operating reports. Note: Extracted March 2017. One month rolling average based on a combination of average daily visitors and page views. Localized Google sites excluded (google.co.in would rank seventh). a. Monthly mobile users. b. Owned by Alibaba, which reports a single aggregated figure. companies’ data (that is, colocation). Data centers vary in sovereignty, insisting that government data be stored in the size, capability, security, and redundancy. A so-called tier country, driving demand for national data centers. Software 1 data center provides basic nonredundant connections parks have also become popular in developing nations as a between computer equipment and may be prone to elec- way to grow their digital economies, and data centers are trical outages, and a tier 4 center has redundant compon- essential for these facilities. Although the operation of a ents, multiple connections between computers, continues to data center does not create many jobs, they are an essential operate during maintenance, and is protected against most platform for companies using them to generate revenue and physical events (Uptime Institute 2012). employment (Dutch Data Center Association 2015). Statistics vary widely about the number of data centers in Telecommunications carriers are particularly keen on data the world. One challenge is that most centers are “small racks centers as a way to offset declining revenue from traditional in computer rooms in smaller companies” (Gartner 2015). voice services. Japan’s biggest carrier, NTT, is one of the largest Although the number of data centers has grown rapidly, data center operators in the world, with more than 140 across growth is forecast to slow because of the trend toward larger the globe.8 Leading carriers have formed a working group of the spaces. More information is available about giant data centers, Open Compute Project for the adoption of common standards referred to as “hyperscale” because of their size and ability to for data centers.9 Operators in developing nations from Paraguay add servers and storage as needed. They are operated by about to the Philippines are busy constructing state-of-the-art data two dozen global IT companies, including heavyweights such centers. Mobile group Millicom has been launching data centers as Amazon, Microsoft, and IBM, as well as enterprises provid- in African countries (Millicom, n.d.), and its Paraguayan data ing cloud-computing services. The 259 hyperscale data centers center won an award in 2016 for its modular design (Millicom in 2015 are projected to grow to 485 by 2020 (figure 2.3). 2016). The Philippines Long Distance Telephone Company has Although the majority of hyperscale data centers are in constructed eight data centers across the country to be close to IT developed nations, data center growth in emerging markets parks and support its cloud-based service offerings (Verge 2015). has ticked up. As more users are connected to the inter- But data centers require significant electricity to power net in lower-income nations, demand for data is rising. and keep equipment cool. According to a study, data centers IP traffic is forecast to grow fastest in developing regions in the United States accounted for 2 percent of that coun- during 2015–20. Some countries are concerned about data try’s electricity consumption in 2014 (Shehabi et al. 2016). 12 Information and Communications for Development 2018 Figure 2.3 Hyperscale data centers 600 47 50 % share of data center servers (installed base) 43 45 500 38 40 33 485 Hyperscale data centers 447 35 400 27 399 30 300 346 25 21 297 20 259 200 15 10 100 5 0 0 2015 2016 2017 2018 2019 2020 Source: Adapted from Cisco Global Cloud Index, 2015–20. The data center industry is therefore constantly looking for cool climate12 and, in 2017, the government launched its ways to reduce reliance on fossil fuels, particularly given the first data center (Moss 2017). possibility of data rationing due to shortages of electricity: Many developing countries face a challenge competing with hyperscale data centers abroad given that electricity If governments and companies decide to rely upon costs tend to be relatively high. In Rwanda, the government increased energy generation, they will not be able is considering subsidizing data center electricity costs to to keep up with the demands of big data without attract more digital companies to the country and for local significantly contributing to environmental pollu- firms to transition their websites to local hosting enterprises tion levels. In this future, how would the world look? (Internet Society 2017). Small island developing states Would governments step in to regulate Facebook generally have high electricity costs due to the absence of usage, only in daylight hours? Would citizens have local energy sources: 8 of the 10 most expensive countries the right to only 12 Google searches per day? Should for electricity are such states. However, they are surrounded we tax companies on their levels of data usage? This by a useful resource: cool seawater. As noted, Google uses might seem laughable now, but data rationing is a seawater to cool its Finnish data center and Microsoft is likely outcome if we do not tackle data growth and testing underwater data centers (Cutler et al. 2017). Maur- the underlying demands placed on power consump- itius has also experimented with ocean water to cool data tion. (The Green Grid 2016) centers (Elahee and Jugoo 2013). Some Pacific small island This has made geographies with cool climates and developing states and other coastal economies have taken abundant hydro or geothermal energy attractive locations advantage of new submarine cable connections to bundle for data centers. Google’s Finnish data center is built in data centers into the landing station, lowering construc- a restored machine hall designed by renowned architect tion costs. Samoa, which has among the highest electricity Alvar Aalto and draws on water from the Bay of Finland costs in the world (figure 2.4), installed an energy efficient for cooling.10 Its data center in West Dublin does not need prefabricated data center in its new cable landing station air conditioning units because of Ireland’s cool climate.11 (Flexenclosure 2017). Many developing nations also have Some developing nations have similar environments, abundant sunshine with great potential for solar energy making them ideal for data centers. The Data Center Servi- to lower costs. This is the thinking of mobile group MTN, ces data center in the Thimpu TechPark draws on moun- which deployed Africa’s first solar data center at its head tainous Bhutan’s abundant hydropower and year-round office in Johannesburg (van Zyl 2014). Supply: Data Connectivity and Capacity 13 Figure 2.4 Price of electricity (US cents per kilowatt-hour) helping to sustain the internet. As the volume of data transmitted over the internet accelerates, IXPs have become 40 even more relevant for ensuring that it is quickly exchanged among different parties. 35 The largest IXPs (measured by traffic or members) are 30 mainly in Europe, with its long tradition of multistakeholder internet cooperation. The biggest is the German Internet 25 Exchange (DE-CIX) founded in 1995, with locations in 20 Dusseldorf, Frankfurt, Hamburg, and Munich. DE-CIX Frankfurt is the world’s leading internet exchange, with peak 15 traffic of 5.6 terabits per second in March 2017. More than 700 networks are connected, and access is available from 10 20 data centers across the city. The networks connected to 5 DE-CIX are a smorgasbord of giant telecommunication carriers (such as Deutsche Telekom, China Telecom, Veri- 0 zon, and NTT) and emerging country operators (such as es ius an ay ca da s a d d te mo lan lan pin Sri Lanka Telecom, Telkom South Africa, and Telkom Indo- gu fri ut an rit ta Sa Bh hA ra au Rw Fin Ire ilip dS Pa M ut Ph ite nesia), big IT firms (such as Apple, Google, and Microsoft), So Un World average and content and service providers (such as eBay and Face- Source: Adapted from World Bank 2017. book). DE-CIX began expanding abroad in 2012 and now Note: Figure provides the latest available data. operates IXPs in Dallas, Dubai, Istanbul, Madrid, Marseille, New York, and Palermo. Reliability is critical for data centers. Developing coun- Although IXPs are burgeoning in most developed markets, tries, particularly in Africa, will need to improve the quality growth has been uneven in developing nations. According to of the electricity supply to create the proper environment to one source, 78 economies are still without an IXP (map 2.1). attract investment in data centers (box 2.1). This will require The establishment of an IXP is often hampered by electricity sector reform and prioritizing reliability for firms. small markets, vested interests, and limited or unbalanced competition. Powerful incumbents with a high level of IXPs and caches: Closer to the edge control over international gateways prefer that ISPs use Although trends suggest a move toward larger data centers, their overseas links for IP transit. Nevertheless, develop- the tendency is toward pushing data closer to the user or ing regions, such as Latin America and Africa, have been the “edge” to reduce latency and lower costs (Leavitt 2010). adding ISPs at relatively high levels (table 2.2). Where no Having data close to end users is critical, particularly in the powerful incumbent exists, IXPs can thrive. This is the financial sector, in which a few milliseconds advantage has a case of the Rwanda IXP, where the historical operator no huge potential impact (Anthony 2012). This is raising traffic longer exists. The IXP has 13 members, including all of the on internet exchange points (IXPs), places where telecom country’s infrastructure-based ISPs. Peak traffic load was carriers and content providers come together to exchange over 1 gigabit per second in March 2017, up more than their traffic (peering). This is cheaper, particularly for devel- 50 percent from the previous year. An IXP is particularly oping countries, since internetwork traffic does not need relevant in landlocked countries like Rwanda, which is far to be sent over costly international links only to return. In from undersea fiber-optic cables. addition, ISPs do not need to make peering agreements with each potential partner. IXPs also improve quality since they Cloud computing: Back to the future are situated closer to the user and hence have less latency. “Soft” benefits are also associated with IXPs, such as devel- The ability to store and process data remotely dates to the oping technical skills and fostering a culture of cooperation, early days of computer networks. Back then, end-user devices 14 Information and Communications for Development 2018 Box 2.1 Sub-Saharan Africa: Reliable electricity and the digital economy Many countries in Sub-Saharan Africa seek to diversify their economies with information and communication technologies (ICT), including expanding ICT as a sector and increasing its use in enterprises. The data center is a core element of ICT infrastructure. These facilities are a vital engine of the digital economy, storing data, hosting websites, and enabling cloud-based applications. Data centers are virtual data factories that make productive use of electricity, with measurable economic impact on gross domestic product, employment, and government tax revenue. Data centers consume lots of electricity to power computer equipment and keep it cool. In 2011, Google reported that it used 260 megawatts of electric power for its data centers, which is greater than the 2014 installed capacity in 19 Sub-Saharan African countries. Data centers require high reliability to ensure seamless, nonstop data flow. Reliability is defined by industry standards, ranging from 99.670 percent availability with no more than 29 hours of interruption per year for tier 1 data centers, to 99.995 percent reliability with just 0.8 hour of interruption per year for the highest, tier 4 centers. Most Sub-Saharan African nations would find it diffi- cult to meet even tier 1 reliability. The standards also call for a guaranteed source of electrical backup that can power the center for at least half a day. Lack of enterprise-grade reliability requirements for industry certification generally rules out the feasibility of large data centers in many Sub-Saharan African countries. Although virtually every country in the region has a data center, the centers are small, serving a narrow set of business and government users. Because of the region’s challenging environment for reliable and inexpensive electricity, most businesses host their data outside the region. This results in a large volume of data transmitted to overseas data centers, requiring significant amounts of international internet bandwidth. Along with connectivity and storage costs, it takes a longer time to access overseas data centers, raising latency. Security is also an issue, as increasing amounts of government, business, and personal information are transmitted abroad, with vague data protection. To build up its national data center industry and improve latency, Rwanda launched an initiative to repatriate 1,000 websites hosted abroad (RICTA 2015). An analysis of the program found that quality was improved for domestic users because of faster access to the sites (Internet Society 2017). Visitor engagement was high, with more page views and return visits due to the enhanced performance. The skills of web-hosting employees increased, due to technical requirements to manage additional websites. Although latency improved, it is still difficult to convince local businesses to place their websites in Rwanda because of the lower price of hosting overseas. This is primarily because of the high cost of electricity for data centers in the country. The government is contemplating subsidizing the cost of electricity for local data centers to make local hosting more attractive, improve latency, and strengthen data sovereignty. Despite concerns about reliability, interest is growing in installing large data centers in the region to achieve better latency and reduce the cost of international bandwidth. In 2017, Microsoft, one of the world’s largest owners of data centers, announced it would build two in South Africa to support its cloud-based services. Notably, South Africa’s electricity supply is considered the second most reliable in the region after Mauritius. The new data centers will be faster than accessing cloud services in Europe or the United States, international connectivity costs will be lower, and trust higher, as the centers will have to comply with South Africa’s data protection law. Electricity reliability is critical for other countries in the region that want to develop their digital economies. Source: World Bank Group 2018. Supply: Data Connectivity and Capacity 15 Map 2.1 Internet exchange points around the world, 2018 Internet exchange points (IXPS) around the world City with: 15 IXPS 7 IXPS 1 IXP 1 IXP 90 IXPS Countries without IXPS IBRD 43803 | SEPTEMBER 2018 Table 2.2 Internet exchange points by region Internet exchange points Domestic bandwidth production February February Net Percent February February Net Percent Region 2016 2017 change change 2016 2017 change change Europe 136 175 +39 +29 35.9T 41.8T +5.84T +16 North America 81 94 +13 +16 3.41T 4.55T +1.14T +33 Asia and Pacific 75 92 +17 +23 2.56T 3.51T +953G +37 Latin America 45 72 +27 +60 2.22T 2.75T +524G +24 Africa 30 42 +12 +40 325G 417G +92.6G +29 Source: Packet Clearing House, internet exchange point directory reports, http://wwww.pch.net/ixpdir/summary. Note: G = gigabytes; T = terabytes. were “dumb” terminals hooked up to large mainframe 7 Mbps in 2010. Average mobile speeds were considerably computers that did all the work. The invention of the slower at 2.0 Mbps in 2015, but with large device differ- personal computer was revolutionary in providing users ences; smartphones averaged 7.5 Mbps around the world with their own device that could run applications and store in 2015 and are forecast to rise to 12.5 Mbps by 2020. data. The process is again reverting to centralized control, where data is increasingly stored and processed over the • Greater storage. Storage available over the cloud is vastly “cloud” on anonymous data servers. Three main reasons superior to what can be saved on a desktop, laptop, or explain this: tablet computer or smartphone. • Faster networks. Rising internet speeds are making the • Proliferation of smart devices. As the number of devices a transfer of data between device and cloud increasingly person owns increases, the cloud provides a useful way transparent. According to Cisco (2017), average global of keeping them all synced. There were 2.2 devices per fixed broadband speeds were 25 Mbps in 2015, up from person worldwide in 2015, projected to rise to 3.4 by 2020. 16 Information and Communications for Development 2018 Several acronyms are used to identify different cloud cyberattacks, such as against IT giant Yahoo, affecting services. Infrastructure as a service (IaaS) offers comput- some 1.5 billion accounts (McGoogan 2017). ing power and storage. Platform as a service (PaaS) offers computer programs and other tools for users to develop Internet of Things: Data is all around their own applications. Software as a service (SaaS) offers complete applications and supporting upgrades and According to Swedish IT equipment manufacturer Ericsson, maintenance. “things” connected to the internet will overtake devices There are a number of benefits for cloud users, including used by humans in 2018 (Ericsson 2015). Cisco reckons reduced need for IT expertise, flexibility for scaling, and some 12 billion devices will be connected to the internet by consistent application rollout and maintenance for large 2020 that talk to other devices or computers, up 20 percent organizations. Free cloud services also exist that provide a year from 2015 (figure 2.5, panel a). These so-called office-like application tools useful for small and medium machine-to-machine connections form the heart of what enterprises (SMEs), as well as social network pages and is referred to as the Internet of Things (IoT), an inter- blogs. This is particularly relevant for developing countries connected ecosystem of sensors, meters, radio frequency where the cost of licensed software can be an obstacle to identification chips, and other gadgets. Traffic from these creating applications and services. things will grow at twice the rate they are being connected, Though cloud computing offers a number of benefits, or 40 percent a year from 2015 to 2020, from 1 exabyte per it comes with costs and risks. Users will utilize more of month to 6.3 (figure 2.5, panel b). their data allowance accessing cloud services, and busi- Machines have talked to each other for years over nesses face migration costs either converting to the cloud communication networks using electronic data interchange or changing cloud providers. Risks include security and and other formats, largely to exchange financial infor- privacy breaches as well as potential loss of service due mation such as transactions from bank automated teller to communications or electrical failures. These risks have machines or companies ordering products or services from been well publicized through headline stories detailing each other. The IoT expands scope, as the things doing the Figure 2.5 Global machine-to-machine connections and traffic a. Connections (billions) b. Traffic (exabytes per month) 14 7 6.3 12.2 12 6 10.4 10 5 4.7 8.7 8 7.2 4 3.5 5.9 6 3 4.9 2.4 4 2 1.6 1.0 2 1 0 0 2015 2016 2017 2018 2019 2020 2015 2016 2017 2018 2019 2020 Source: Cisco Visual Networking Index Global IP Traffic Forecast, 2015–20. Supply: Data Connectivity and Capacity 17 Figure 2.6 Machine-to-machine connections per 100 people, OECD member countries, June 2016 Millions 77.3 30 12 10 20 8 6 10 4 2 0 0 Ze n ite Ire e ep p. A u ce ds Es tes Fra ia L J m ov mb n Ko Spa c Cz re in Ice lic Ch a Gr ile M ria ico Re urg Ne Finl y Un Den aly d S rk Sl key No nd th and Sw Pol l itz and Tu nd ing nd Po nd bli de nc Sl uxe apa i ga a h R Re n do en ite ma ee lan ub rw st ex ta to ala la d K la la It ak o pu e r rtu ov Sw ec a, er er w Ne Un M2M connections, per 100 inhabitants (left scale) M2M connections, millions (right scale) Source: Organisation for Economic Co-operation and Development (OECD). Note: M2M = machine-to-machine. communicating are generally small devices and sensors Many analysts see fifth-generation (5G) wireless tech- tracking everything from utility use to automobile move- nology as critical for the IoT because of its expanded data ments. handling. According to one report, 5G networks can process A report from the International Telecommunication about a thousand times more data than today’s systems Union (ITU) and Cisco argues that the IoT could be bene- (Hellemans 2015). Of particular relevance for the IoT is ficial for developing countries,13 since it lowers the cost of 5G’s ability to connect many more devices (such as sensors service monitoring and delivery, allowing countries to gain and smart devices) than previous generations of wireless. A in areas such as health and energy over a shorter timeframe major milestone was reached in December 2017 when the than ever before. The ITU has formed a study group to first 5G specification was approved by the 3rd Generation enhance global standardization and collaboration on the Partnership Project and endorsed by many of the world’s IoT.14 One example is Ghana, where sensors are helping leading telecommunication equipment manufacturers and improve the vaccine supply by indicating whether refrigera- operators (3GPP 2017). The ITU’s World Radiocommuni- tion was affected during transport. cation Conference 2019 aims to establish standards for The gap in IoT adoption is wide, according to statistics. spectrum management and harmonization for 5G.15 Some For example, adoption of machine-to-machine communi- countries cannot wait, such as the United Kingdom, which cations, a subset of the IoT, varies tremendously in earmarked some 3.4 GHz for 5G and auctioned it in March Organisation for Economic Co-operation and Develop- 2018 (Ofcom 2018). Operators in several countries have ment (OECD) countries, with Sweden and New Zealand announced commercial deployments of 5G before the end ahead by some margin in respect to machine-to-machine of 2018. connections per 100 people (figure 2.6). Sweden’s tele- communication companies are striving to be leaders in Data-driven business models Internet-of-Things services, and the country also has a sizable Internet-of-Things startup ecosystem (Luleå This section looks at how access to the internet and its University of Technology 2014). And a major reason for content is priced and which parties earn the revenue. It New Zealand’s high penetration is that the main gas and examines the different ways for earning revenues from electricity company has installed more than a million network access and content, including subscriptions (post- smart meters in homes and businesses across the country paid and prepaid), advertising, and transaction fees. It looks (Vector 2017). at the consequences of data-driven business models for the 18 Information and Communications for Development 2018 traditional pricing approach used by telecommunication The third, and most predominant, particularly in devel- operators, with potential impacts on network investment, oping countries, is prepaid, in which the price is tied to market concentration, and net neutrality. a fixed amount of data. Prepaid must generally be used In reviewing alternative pricing structures, the differ- within a specific time, but does allow for flexibility in ence between access to the internet and to its content that a small amount of data can be purchased for a day varies in their financial significance. Users typically pay or a week. for access to the internet based on the volume of data they To justify investment in higher-bandwidth infra- use, whereas content is not generally priced by volume. structure, a number of telecom operators around the While a user may view free video streaming services that world have diversified into providing video services, generate large amounts of traffic, an online shopping enabling them to provide so-called triple-play offers transaction generates little data traffic but, in aggregate, (such as voice, data, and video). This has led to bundled creates significant value for the seller. This contradiction offers in which all three services are offered together at poses a significant challenge for measuring the economics a price higher than that for buying any single service of data: alone, but at a discount compared with purchasing each service separately. Ironically, although traditional broad- The great challenge for economic measurement cast operators moved to protect their markets through stems from the fact that the consumption of digital legal challenges, neither they nor the telecom operators products often does not involve a monetary trans- foresaw the greater threat looming from streaming video action that corresponds to its value to consumers. services. Digital products delivered at a zero price, for The challenge today is that data traffic is exploding instance, are entirely excluded from GDP (gross while the traditional revenue earner for telecom operators, domestic product), in accordance with the inter- voice and text, is declining as users shift to over-the-top nationally-agreed statistical standards. . . . The gap (OTT) services delivered over the internet. Data traffic from between what is measured and what is valued grows smartphones surpassed voice traffic in 2011 and since then every time access is gained to a completely new good has grown at a tremendous rate, accounting for 96 percent or service or when existing goods or services are of mobile traffic (figure 2.7, panel a). However, operator offered free as is often the case after digitalisation. revenues have not kept pace. In 2015, voice still accounted The question is how these new forms of consump- for almost half (46 percent) of telecom operator revenue tion should be accounted for in economic statistics (figure 2.7, panel b), down sharply from two-thirds of oper- such as GDP. (Bean 2016, 74) ator revenue in 2010. Total telecom revenue has only been Consumer pricing at telecom network operators growing at 2 percent a year. has evolved with technological change, from price per minute, to price per speed, to price per data consumed. How to pay: Advertising-funded versus Before the internet, voice was king and operators gener- subscriber-funded networks ally charged for it on a per-minute, metered basis. Before The internet has largely been characterized by “free” content. broadband, consumers mainly accessed the internet Although some sites, mainly news and streaming audio using dial-up connections with pricing as if it was a voice and video, charge subscriptions, the most popular websites call (that is, minutes of use). The emergence of fixed are free. Most of these sites earn revenues from advertising broadband led to pricing by speed. With mobile broad- through a two-sided market strategy of providing users with band, it is harder to guarantee speeds because of different incentives to join their platform (Dallas 2014). As the inter- coverage ranges and a variety of handsets. Three main net audience has grown, firms are increasingly flocking to models exist for pricing data for mobile networks. One is place ads on the web and smartphone apps. flat rate, in which data usage is not metered but there is a The growth of digital advertising is astonishing. Global fine-print data cap that, if surpassed, results in the speed internet advertising accounted for more than a third of being lowered. The second is postpaid, in which a certain advertising expenditure in 2016, slightly behind television amount of data is included with a monthly subscription. (figure 2.8, panel a).16 Supply: Data Connectivity and Capacity 19 Figure 2.7 Global network traffic and retail telecom revenue, selected countries a. Global network traffic b. Retail telecom revenue 100,000 100 700 17 16 14 90 21 19 24 600 80 596 10,000 567 576 577 538 550 500 70 32 35 37 60 39 41 400 43 1,000 50 22 300 40 19 17 16 30 15 200 100 14 20 29 31 32 100 22 26 10 19 10 0 0 2010 2011 2012 2013 2014 2015 11 10 13 16 18 19 20 12 14 15 17 21 22 20 20 20 20 20 20 20 20 20 20 20 20 20 Fixed voice Mobile voice Fixed broadband Voice traffic (petabytes/month) Mobile data Total (billion pounds sterling) Smartphone data traffic (terabytes/month) Sources: Adapted from Ericsson 2016 and Ofcom 2016a. Note: Panel a has a logarithmic scale. Panel b is based on 18 major economies: Australia, Brazil, China, France, Germany, India, Italy, Japan, the Netherlands, Nigeria, Poland, the Republic of Korea, the Russian Federation, Singapore, Spain, Sweden, the United Kingdom, and the United States. Digital advertising spending is highly concentrated, decision, the U.S. government decided in 2017 that ISPs with the vast majority going to two companies: Google can sell browsing histories to advertisers without the user’s and Facebook, who between them received US$106 consent (Kelly and McLean 2017). billion in internet ad revenue in 2016, 64 percent of the The concentration of advertising spending reinforces global total (figure 2.8, panel b). Their share is growing, the large properties at the expense of the millions of up 20 percentage points since 2010. They both have huge smaller ones. It also threatens infrastructure investment, user bases, with each reporting more than a billion users since the large websites sit on top of telecommunication of their services, which is attractive to advertisers. But networks, yet the networks do not necessarily receive reasons for placing ads on them differ, partly explaining advertising revenues. Although there are payments from why advertisers often put ads on both. According to one content businesses to carriers for bandwidth, it is not clear digital ad analyst, “Facebook believes the most impor- that they make up for the huge amount of traffic gener- tant thing is identity in ensuring ad effectiveness . . . they ated. At the same time, if not transparent, the payments know who you are and so much about you” whereas between content providers and telecommunication carri- “Google believes identity is secondary to intent. What’s ers pose net neutrality concerns, since a payment may important is what you want right now because adver- imply an enhanced traffic service, to the detriment of tising products and services fulfils a want or need” other content providers. (Garrahan 2016). Another concern is that as a few global sites thrive Digital advertising is causing a huge transfer of wealth they are branching into areas they have no expertise in from traditional advertising outlets (such as newspapers and and for which their automated content controls pose radio) to internet companies. Telecom network operators challenges. One area is news, with traditional news have also not largely benefited from the value advertisers outlets hit hard by a rapid decline in advertising. The rise place on internet content. However, in a highly contested of news on IT content sites is of particular concern, with 20 Information and Communications for Development 2018 Figure 2.8 Global advertising revenue a. Distribution of global advertising spending (percent) b. Percentage of internet advertising revenue 100 600 100 90 90 500 36 80 80 42 41 48 46 45 70 70 56 15 16 18 20 400 24 30 34 60 60 16 5 7 9 12 50 300 50 4 3 40 40 200 30 30 48 49 48 49 47 48 20 20 42 100 10 10 0 0 0 2010 2011 2012 2013 2014 2015 2016 2010 2012 2014 2016 Cinema Outdoor Radio Magazines Others Facebook Google Internet Newspapers Television Total (US$ billion; right scale) Source: Adapted from Zenith Media, various years. objectivity and facts increasingly called into doubt given Besides advertising, some internet companies charge the explosion of sources (Garret 2016). At the same time, subscription fees, particularly those involved in content. legitimate news and information is sometimes blocked, One example is Netflix, which offers streaming television illustrating the weaknesses of robotic software agents and film delivered over the internet. E-commerce sites such trying to determine what is appropriate. For example, as Amazon, eBay, Rakuten, and Alibaba earn revenue in Facebook blocked a 1972 Pulitzer Prize winning photo two ways. One is as a normal retailer, charging a markup of a Vietnamese girl over concerns about nudity (Scott over price. The second is when the platform is provided to and Isaacs 2016). The company was accused of abus- third-party sellers, in which case the platform owner earns ing its power and the photo was later reinstated (Wong transaction fees. 2016). Google has been under fire for placing ads next Table 2.3 contrasts average revenue per user per month to extremist content on its YouTube video-sharing site for the various internet payment models and for large (BBC 2017). These examples have led to a growing internet-based companies. Prices have also been converted argument that IT firms posting news stories should be to purchasing power parity to adjust for differences in the subject to regulations similar to those that media firms cost of living. Subscription-based video viewing has the face (Ingram 2017). highest average revenue per user, and advertising on social Advertising does support free services for users, which is media the lowest (but more subscribers). The telecom particularly attractive for developing countries. These services carrier in the list has the second highest average revenue per include email, office applications, storage, and maps delivered user (purchasing power parity). over the cloud (Greengard 2010). The availability of free and Some content companies are becoming involved in devel- legal cloud-based applications and services is putting a dent in oping networks, feeling that infrastructure development is software piracy, which has been dropping, and the focus has not keeping up with the vast growth in data traffic. A desire shifted from loss of revenues to the cybersecurity dangers of to capture more users from developing countries by making unlicensed software (BSA The Software Alliance 2016). it easier for them to get online is also driving this effort. Supply: Data Connectivity and Capacity 21 Table 2.3 Average revenue per user from internet data, 2016 Average revenue per user (month) US At purchasing Company (Country) Revenue model dollars power parity Users (millions) Note Facebook Advertising 2.63 2.63 1,860 World’s largest social network platform (United States) Alibaba (China) Retail margin/ 2.37 4.47 423 World’s largest business-to-consumer retailer transaction fees (gross merchandise value) China Mobile (China) Subscription and usage 4.31 8.14 849 World’s largest mobile operator Netflix Subscription 8.61 8.61 94 World’s largest paid video streaming service (United States) Source: Adapted from operating reports of companies shown in table. Facebook and Google are investing in a variety of communi- • Viber offers calling, video, and messaging services to cation ventures. more than 800 million people and is owned by Japanese At Google, this includes providing fiber internet access e-commerce company Rakuten.23 in several U.S. cities,17 offering fiber backbones and • Skype, owned by Microsoft, offers calling, video, and Wi-Fi in Ghana and Uganda,18 using hot air balloons to messaging services. extend internet connectivity,19 and supporting the use of white spaces for making more spectrum available.20 It also • Netflix provides paid video to 94 million subscribers (as controls Android, the leading mobile phone operating of the end of 2016) around the world. It competes with system.21 telecommunication service providers that also offer video Facebook has been looking at satellites, drones, and (such as IPTV). However, the bigger impact on the tele- solar-powered airplanes for extending internet access coms is the volume of traffic Netflix generates. (O’Brien 2016), and is getting involved in networking gear The impact of OTT has been particularly strong in (Bort 2016). The growing involvement in data transmission what has long been a traditional profit center for tele- by large content firms raises questions about the separa- communication carriers: international voice telephone tion of carriage and content, with the possibility of a few calls. By 2012, Skype was already handling a third of inter- companies dominating both internet content and access. national telephone traffic (Goldstein 2013). By 2016, OTT traffic using voice over internet protocol exceeded inter- The rise of “over-the-top” service providers national traffic provided by telecommunication operators OTT refers to data services provided over telecommunication (figure 2.9). networks. The impact of OTT on telecommunication oper- Similarly, telecom carriers have seen a sharp fall in ators comes from either competing with traditional revenue conventional messaging services (such as SMS) used on their sources, such as voice calls and messaging, or depositing a networks. In 2015, an analysis of 17 countries found that the lot of traffic on their networks. A lack of clear metrics makes number of messages declined in 14 of them (figure 2.10). understanding the OTT market a challenge. Some of the The danger is that if carriers cannot offset the loss of most popular OTT services have been purchased by larger revenues, less money may be available for future infra- companies that do not provide disaggregated financial or structure investments to handle the rapid increases in data operating indicators. Nevertheless, considerable circumstan- traffic. The ITU has created a study group to examine tial evidence suggests the impact is significant. this issue.24 The West African telecommunication operator Notable OTT providers include the following: Sonatel paints a dire picture, estimating that between 2016 • WhatsApp, purchased by Facebook in 2014, offers messa- and 2020, its losses from OTT in the international segment ging and calling services and claims to have about a will be CFA francs 256 billion (US$432 million) in Senegal, billion users in more than 180 countries.22 CFAF 164 billion (US$277 million) in Mali, CFAF 79 billion 22 Information and Communications for Development 2018 Figure 2.9 International carrier and over-the-top traffic (billions of minutes) 800 727 700 600 568 547 556 553 517 530 514 500 473 431 418 405 383 400 350 308 304 300 274 237 214 200 160 116 85 100 55 22 33 8 14 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Carrier Over-the-top Source: Telegeography. Figure 2.10 Mobile messages, year-over-year change (percent), 2014–15 10 6.3 5 2.5 0.8 0 –1.1 –0.1 –5 –6 –5.8 –5.6 –10 –8.6 –7.9 –15 –12.1 –20 –16.7 –18.8 –25 –24.9 –30 –27.7 –30.7 –35 –40 –36.8 ny re ds ina e m n ain ia ia en es a il ly p. d nc tio ali po az lan do Ind r ma lan Ita at Re ed ge Ch Sp Fra str a Br a ng St Po Sw r er er Ni ng a, Ge Au Ki th ed ed re Si Ne ed Ko nF it Un it ia Un ss Ru Source: Adapted from Ofcom 2016b. (US$132 million) in Guinea, and CFAF 12 billion (US$20 are developing their own OTT products. Others are million) in Guinea Bissau.25 It also estimates that taxes paid including large bundles of their own offerings, such as to the government and dividends for its shareholders will fall free calls or text in packages. Many are diversifying into CFAF 243 billion (US$410 million). opportunities in areas such as cloud computing, IoT, and Telecom operators have developed several responses mobile money. Some operators are trying to do all of the to OTT. They have argued for regulating OTT provid- above. Also relevant is taxation for OTT firms without a ers that provide voice and text services in the same way physical presence in the country in which they are provid- that telecom operators are regulated. Some operators ing service. Though they may offer the service for “free,” Supply: Data Connectivity and Capacity 23 digital advertising revenues often subsidize this. The lack The promise and perils of zero-rated services of OTT taxation in most developing countries gives them Many internet applications are based on the so-called a cost-structure advantage over domestic telecommunica- freemium model, in which consumers get basic features tion operators. at no cost and can access premium functionality for a Some OTT services compete with traditional tele- subscription fee. However, users must still pay for the data communication services such as voice and text. However, consumed using these applications. Zero-rated services others provide a different challenge, particularly OTT provide access to certain content without it applying to a video providers. These include those that provide tele- user’s data cap. Some firms with desirable content, such as vision and films through subscription services such as Facebook, have worked with operators, mainly in develop- Netflix as well as free services such as YouTube and ing countries, to provide access to their services without Facebook, in which video posts are increasing rapidly its affecting a user’s data allowance. At the same time, (Peterson 2015). They not only compete with telecom some operators are striking agreements to provide access operators that provide video services, but are also respon- to some services free to their customers. For example, sible for a substantial portion of traffic going over the T-Mobile in the United States does not charge data usage networks. Netflix, YouTube, and Facebook are among the for customers on certain subscription bundles when top traffic applications in most regions, comprising 42 they stream music.26 Other operators provide discounted percent of fixed-access and 34 percent of mobile-access, access to a bundle of social networking services, but this peak-period traffic (table 2.4). When all of the properties is technically not zero rated, since users still pay for data of Facebook (that is, Instagram, WhatsApp, and so on) access but at a discounted rate. and of Google (YouTube, Google Cloud, Google Market, It is argued that zero-rated services provide a taste of a and so on) are considered, they account for an even larger slimmed-down version of a service and users will eventually share. The two account for more than 60 percent of total pay for access to the full internet (Facebook: Internet.org traffic on Latin American mobile networks, for instance 2015). Text-only versions may also be relevant for users who (Sandvine 2015). do not have access to high-speed mobile internet and rely Table 2.4 Percentage of aggregate peak period traffic by region, 2015–16 Simple Africa Asia and Pacific Europe Latin America Middle East North America average Fixed access YouTube 16 27 21 30 NA 15 22 Netflix 4 6 34 15 Facebook 9 3 7 6 NA 3 6 Other (top 10) 44 45 38 41 NA 22 38 Other 31 24 29 17 NA 27 26 Top 10 69 76 71 83 NA 73 74 Mobile access YouTube 10 17 20 20 21 19 18 Netflix 2 4 3 Facebook 6 8 16 26 10 16 14 Other (top 10) 49 36 38 34 43 35 39 Other 35 39 26 19 26 26 29 Top 10 65 61 74 81 74 74 71 Source: Sandvine 2015. Note: NA = data not available; blank cells indicate that the service in question is not offered in a particular area. 24 Information and Communications for Development 2018 on slow 2G connections. One example is Facebook Zero, Importantly, cost is not the only or even, in many coun- launched in 2010, providing access to a text-only version of tries, the main barrier to internet access. Users often cite the service. It has now been renamed the Free Basics service, reasons such as no need or lack of skills as the reason they available in more than 60 economies (Facebook: Internet. do not use the internet. In Thailand, 97 percent of those who org, n.d.), with about 40 million users (Constine 2016). In do not use the internet said the main reason was because addition to Facebook, access is provided to other websites, they did not know how to use it or it was unnecessary or a such as Wikipedia, Accuweather, and Bing, as well as to local waste of time (National Statistical Office, Thailand 2016). social impact sites for health and employment. In Brazil, 70 percent of those not using the internet cited a A variation on zero-rated services is sponsored data, lack of interest as a reason.29 So it is not clear to what extent in which companies pay for data usage if a user agrees to zero-rated services get new users online. As one report puts receive ads. Sponsored data is also used for companies to it: “Even with a zero-rated service, the user must still have a pay for their employees’ mobile data usage for work. AT&T, device and an active account with the operator that offers a U.S. mobile operator, offers sponsored data in which users the zero-rated service. This raises the question of whether are not charged for usage if they access a sponsor’s site.27 zero-rated services can bring people online who had not Syntonic, one of the AT&T sponsors, notes: “Millions of previously used the internet” (Alliance for Affordable Inter- prepaid consumers ration their data, impeding discovery net 2016). The report looked at the impact of mobile and exploration of mobile apps and content” and is expand- data apps across eight developing countries, finding that ing its product reach outside the United States to southeast 88 percent of users had already accessed the internet before Asia, India, and Mexico (Syntonic 2017). using a zero-rated plan. This suggests that digital literacy While providing free content seems commendable, only challenges are arguably more important than affordability making certain content available runs contrary to the net to get more people online. neutrality principle of the internet. In February 2016, the Telecom Regulatory Authority of India issued a regulation Data holes: Filling the gaps prohibiting the use of what it called discriminatory tariffs for data services. The authority formed its decision based This section explains why data location, language, and limits on “the principles of Net Neutrality seeking to ensure that are becoming more important than plain access. It also consumers get unhindered and nondiscriminatory access to explores links between data and economic development. the internet. These Regulations intend to make data tariffs for access to the internet to be content agnostic” (Telecom From access to usage Regulatory Authority of India 2016). Several other countries Significant attention has been devoted to the uneven distri- have also banned zero-rated services. On the other hand, the bution of access to information and communication tech- United States Federal Communications Commission struck nology (ICT). As mobile phone penetration rises and access down net neutrality provisions in December 2017. The to the internet increases, the access gap is shrinking. Nine in 3–2 vote by commissioners was along political lines in an ten people around the world were covered by a 2G cellphone allegedly “contentious and messy” public comment process signal in 2016 and 65 percent by 3G; by 2020 these figures (Kastrenakes 2017). ISPs in that country are now allowed are forecast to rise to 95 percent and more than 90 percent, to block, throttle, and prioritize content. About 20 states respectively. The unconnected are increasingly those who are have filed lawsuits against the ruling, and the United States not interested in using or do not know how to use the inter- Congress is considering overturning the ruling if it can net rather than those who have no access or cannot afford muster the votes (Fiegerman 2018). to pay. Investment in digital literacy training is becoming as At the same time, it is argued that zero-rated services important as infrastructure. give an advantage to large companies to the detriment of The geography of data creation, distribution, and use is new startups. As one report notes: “Ironically, if zero-rated lopsided, resulting in a new global data divide. One mani- services were available when large internet companies festation of this gap is content concentration. More than were startups, it is unlikely they would have scaled to the half of the world’s websites are in English (figure 2.11, size they are now.”28 panel b), yet only 984 million people speak English as a first Supply: Data Connectivity and Capacity 25 Figure 2.11 Global internet protocol traffic and websites by language a. Distribution of global internet protocol b. Percentage of websites by language, March 2017 traffic and population, 2015 (percent) 100 3 Other, 6 14% 90 19 Chinese, 80 23 2% 9 Italian, 70 2% Portuguese, 12 3% 60 French, 5 4% English, 50 34 52% Spanish, 5% 40 German, 30 7% 55 Japanese, 20 6% 34 Russian, 10 7% 0 IP traffic share Population share Middle East and Africa Latin America Europe North America Asia Pacific Sources: Cisco; W3Techs; and World Bank. or second language (Simons and Fenning 2018), 13 percent Data usage is driven by factors such as coverage and of the Earth’s population. Another manifestation of the data device, with pricing a major influence. Data pricing divide is from where it flows. Just over one-third of IP traffic varies significantly throughout the world, measured by is generated by North America, with only 5 percent of the the metric of price per GB per month or for compar- world’s population (figure 2.11, panel a). On the other hand, ability, US$/GB (figure 2.13, panel a). In absolute terms, the Middle East and Africa, home to 19 percent of the world’s average price ranges from US$5 per GB per month in population, only generate 3 percent of global IP traffic. South Asia to US$28 in high-income OECD nations. However, in relative terms, high-income OECD nations New metrics of the data age have the cheapest prices (0.9 percent of GDP per capita) Data holes are reflected by uneven data consumption compared with 12 percent in Sub-Saharan Africa. Prices across communities, regions, and nations. The amount of vary significantly in Sub-Saharan Africa, with relative data data used per smartphone—measured as gigabytes of data prices ranging from a little over 1 percent of gross national per month or GB/user—varies tremendously. Smartphone income per capita in Mauritius to 45 percent in Zimbabwe users in North America consumed almost four times more (figure 2.13, panel b). data than those in the Middle East and Africa (figure 2.12, Being data starved is a constraint when it comes to rich panel a). Average global use is forecast to grow more than multimedia educational, health, and livelihood content. fivefold between 2016 and 2022, from 1.9 GB per month However, many useful activities require just narrowband: to 11. Within North America, U.S. mobile broadband a quick e-commerce transaction, a text message to check subscribers use more than 1.8 times as much data as their produce prices, or a phone call in an emergency. Hourly, neighbor to the north, Canada, and 3.6 times more than daily, and weekly prepaid options can enhance affordability their neighbor to the south, Mexico (figure 2.12, panel b). in these circumstances. 26 Information and Communications for Development 2018 Figure 2.12 Mobile data usage a. Data traffic per smartphone b. Mobile data usage per mobile broadband (gigabytes per user per month) subscription, 2016 (gigabytes) 3.0 North America 25 2.7 5.1 2.5 Western Europe 22 2.7 Central and Eastern Europe 15 2.0 1.9 11 1.5 World 1.5 1.9 Latin America 9.6 1.6 1.0 0.7 Asia and Pacific 9.5 1.7 0.5 Middle East and Africa 7.6 1.3 0 0 5 10 15 20 25 30 United States Canada Mexico GB/user/month 2022 2016 Sources: Adapted from Ericsson 2016 and OECD 2017. Data and economic development and narrowband 2G applications such as text messag- Could the data divide be affecting economic growth in devel- ing or mobile money, which do not use much data. For oping nations? Various studies have looked at the impact of example, a study of grain markets in Niger found that ICTs on economic growth. As businesses and consumers prices dropped 3 percent after the introduction of mobile obtain more high-speed connectivity, they have realized phones because of better access to market information (Aker important benefits in terms of efficiency, new businesses 2010). A study analyzing the economic impact of mobile models, market information, and so on. Some research has money in Kenya found its use decreases prices of competing focused on the impact of data on the economy. Four studies money transfer services and increases levels of financial looking at public sector open data found impacts ranging inclusion (Mbiti and Weil 2011). An econometric analysis from 0.4 percent to 4.1 percent of GDP (ODIHQ 2015). on the impact of telecommunications in Senegal found A European Parliament report states that big data and the no statistically significant effect from broadband; on the data-driven economy will bring 1.9 percent in additional other hand, plain mobile communications had a significant GDP growth by 2020 (European Parliament 2016). contribution, with each percentage point increase in mobile A Deloitte study suggests that data usage affects economic penetration contributing 0.05 percent to GDP (Katz and growth (Williams et al. 2016). Based on mobile data usage for Koutroumpis 2012). These findings suggest that the data 14 countries between 2005 and 2010, the study found that a nuggets are small, often lost in the sea of video and social doubling of mobile data consumption added 0.5 percentage media traffic, and sometimes not even transmitted over the point to GDP growth a year. internet. While the study suggests an econometric link between data consumption and economic growth, the exact reasons Conclusions: Toward sustainable seem fuzzy. It is puzzling, given that most internet traffic is national data ecosystems video entertainment, which is not likely to have a tremen- dous economic effect. Other studies suggest that it may not As the universe heads inexorably into the data era, there are be the quantity of data that is important, but rather the winners and losers. Consumers have access to “free” content value of the data. In many developing countries, economic and services in exchange for their personal information and impacts have been noted from basic cell phone voice calls time spent posting information (Thornhill 2018). It has also Supply: Data Connectivity and Capacity 27 Figure 2.13 Mobile data pricing a. 2015 30 28 14 11.8 25 12 18 10 20 17 16 8 15 13 14 6 10 4.0 6 4.7 4 5 2.4 5 0.9 2 1.1 2.0 2.2 0 0 ia ific sia ca a ea d tri D ld ric ibb n As un C fri or ar ca a ac hA co OE n es Af nA W l dP ra ut er me th e C ri nt ra an th Ame So or Ce mb co ha dN ia me h-in Sa nd As tin an b- ea g La st Su Hi st Ea p Ea ro Eu le idd M Price of 1 gigabyte (US$, left scale) Price as a percentage of GDP per capita (right scale) b. 2016 40 50 35 45 40 30 35 25 30 20 25 15 20 15 10 10 5 5 0 0 Ta ana Bo nya Ni us M mer e e da e Rw ue bia a Ye ia Ug so Gh a d’I a Et en rki pia Su a am n Za a mb nin Na n Se ru ut ria Ja gal M li ba i Ca ir Th bw an ic i c oz oo d da Zim law mb Cô nzan a Pe i Fa fri an biq m an o So ge rit ma mi Bu hio Ga Be M Ke tsw ne v ia, hA au a na M te Price of 1 gigabyte (US$, left scale) Price as a percentage of gross national income per capita (right scale) Sources: Alliance for Affordable Internet; and International Telecommunication Union. Note: OECD = Organisation for Economic Co-operation and Development. been great for content platforms that get free user-generated The globalized nature of the internet is its beauty, but also content and even more valuable, their personal information its peril. Users can access content from Argentina to Zambia that is sold to digital advertisers and data analytics firms. using free platforms to store data and run applications. SMEs On the other hand, the emerging data economy is requiring in developing countries have benefited from free online tools adjustments for telecommunication carriers who are finding and global platforms to increase their visibility. This “free it difficult to fund investment needed for rising data use. The lunch” has resulted in just two U.S. companies—Google and money they are making from data access has not offset fall- Facebook—dominating the platforms by which many of the ing revenue from traditional sources due to OTT. Govern- world’s internet users interact, earning the majority of online ments are finding it increasingly challenging to deal with ad revenue, controlling vast amounts of personal data, and concerns such as net neutrality, privacy, computer crime, generating much of the traffic. They are also extending their and false or incendiary information on the internet, as well operations horizontally and vertically, from online shopping as automated platform censorship. to the provision of telecommunication services. This has 28 Information and Communications for Development 2018 made some governments anxious about the power a few U.S. developing countries can analyze it and put it to good use. companies wield over the internet: And rather than digital advertisers using personal data to sell something, developing country digital scientists can use data Several recent attempts have been proposed by to pinpoint the locations of those living in poverty to better other countries such as Brazil, Germany, China, target assistance (Bohannon 2016). and Russia to better regulate their data sovereignty In short, developing nations need to leverage data to drive requirements against the domination of the US development through locally relevant content and a thriving communications infrastructure and services. These employment-generating digital ecosystem. This will require technical proposals are national email, localised better understanding of data’s potential, investment in core routing of internet traffic, undersea fiber optic cable infrastructure such as data centers and IXPs, and develop- and localised data centre (Nugraha, Kautsarina, and ment of data-driven development applications and services. Sastrosubroto 2015). The rise of dominant internet platforms affects the Notes development of national data ecosystems. Developing coun- 1. According to Cisco, “There is a strong correlation between try telecom operators not only pay for a physical link abroad, experienced speeds and number of video minutes viewed per they also need to pay for data traffic to be exchanged for trans- viewer. As speeds increase ... the number of video minutes mission to an overseas hyperscale data center. Local digital per viewer also increases.” For more, see http://www.cisco businesses struggle to develop new applications and services .com/c/en/us/solutions/collateral/service-provider/visual already dominated by a few free platforms that have achieved -networking-index-vni/vni-hyperconnectivity-wp.html. giant scale because of network effects. The development of 2. See World Wide Web Size at http://www.worldwidewebsize local internet infrastructure facilities such as IXPs and data .com. centers suffers, since so much content resides abroad. 3. Note that fixed and mobile traffic can also go over managed The challenge for many lower-income countries is how IP networks in the case of business or government closed networks. to develop a relevant and sustainable data ecosystem in the 4. British Broadcasting Corporation, BBC at a Glance. https:// current environment. Much of the data consumed around advertising.bbcworldwide.com/brands/bbccom/. the world is entertainment oriented. Yet governments need 5. See the Metro Network “O Globo – Brazil” at TMN Worldwide. development-oriented data to enhance social and economic http://www.tmnww.com/premium-network-global-reach growth. /premium-latam/premium-brazil-rio. A starting point is boosting linguistically and contextually 6. From the Estate of Corinne & Marshall McLuhan, “Commonly relevant local content. This needs to be accompanied by Asked Questions (and Answers),” viewable at http://www investment in infrastructure such as fiber-optic backbones .marshallmcluhan.com/common-questions/. and data centers to bring data closer to users. Applications 7. According to the China Internet Network Information and services that can enhance health and education, such as Center, there were 710 million internet users age 6 and telemedicine and online learning, need to be implemented older in June 2016 (see CNNIC 2016). According to the rather than talked about. National infrastructure deployment United States National Telecommunications and Informa- tion Administration, there were 227 million internet users and the take-up of local content and services can be encour- age 3 and older in July 2015 (See “Digital Nation Explorer” aged by taking a page from mobile tariff structures. Access to at https://www.ntia.doc.gov/data/digital -nation-data-ex locally hosted sites and services such as e-government can be plorer). stimulated with a low “on-net” internet access price, particu- 8. See http://www.ntt.com/en/services/data-center/nexcenter.html. larly since access to locally stored content is cheaper than 9. For more information, see http://www.opencompute.org content stored overseas. Furthermore, digital literacy has to /projects/telco/. be boosted so that taxi drivers can use GPS and not just play 10. For more information, see https://www.google.com/about smartphone games, and SMEs need to move from streaming /datacenters/inside/locations/hamina/. music in their shops to using e-commerce. Digital scientist 11. See https://www.google.com/about/datacenters/inside/locations skills are needed, so instead of being overwhelmed by data, /dublin/. Supply: Data Connectivity and Capacity 29 12. For more information, see http://www.dcs.bt/index.html. BBC (British Broadcasting Corporation). 2017. “Google Apolo- 13. See http://www.itu.int/net/pressoffice/press_releases/2016/02 gises after Ads Appear Next to Extremist Content.” March 20. .aspx#.WNVwdBjMzYI. https://www.bbc.com/news/business-39325916. 14. See http://www.itu.int/net/pressoffice/press_releases/2015/22 Bean, Charles. 2016. “Independent Review of UK Economic .aspx#.WNVxbxjMzYI. Statistics.” Final Report of the Independent Review of UK Economic Statistics (to the Government of the United 15. See https://www.itu.int/en/mediacentre/Pages/2017-PR66.aspx. Kingdom), led by Professor Sir Charles Bean of the London 16. For more on the shift and drying up of print advertising School of Economics. https://www.gov.uk/government/ revenues and move to digital advertising, see WAN-IFRA, publications /independent-review-of-uk-economic-statis- 2016, World Press Trends. http://anp.cl/wp-content tics-final-report. /uploads/2017/02/WAN-IFRA_WPT_2016_3.pdf. Bohannon, John. 2016. “Satellite Images Can Map Poverty.” Science, 17. See https://fiber.google.com/about/. August 18. http://www.sciencemag.org/news/2016/08/satellite 18. See https://www.google.com/get/projectlink/. -images-can-map-poverty. 19. See https://x.company/loon/. Bort, Julie. 2016. “Now Facebook Plans to Eat the $500 Billion 20. See https://www.google.com/get/spectrumdatabase/. Telecom Equipment Market.” Business Insider, November 1. http://www.businessinsider.com/facebook-voyager-optical 21. See https://www.android.com. -switch-telecom-infra-project-2016-11. 22. See https://www.whatsapp.com/about/. BSA The Software Alliance. 2016. “Seizing Opportunity through 23. See https://www.viber.com/en/about. License Compliance.” http://globalstudy.bsa.org/2016/index 24. See http://www.itu.int/en/ITU-T/studygroups/2013–2016/03 .html. /Pages/ott.aspx. Cisco. 2017. “Cisco Visual Networking Index: Forecast and 25. See http://wholesalesolutions.orange.com/content Methodology, 2016–2021.” https://www.cisco.com/c/en/us /download/47961/1369504/version/1/file/OTT+in+Senegal /solutions/collateral/service-provider/visual-networking _Birago+Beye.pdf. -index-vni/complete-white-paper-c11-481360.pdf. 26. See https://www.t-mobile.com/offer/free-music-streaming.html. CNNIC (China Internet Network Information Center). 2016. 27. See https://www.att.com/att/sponsoreddata/en/index.html. “Statistical Report on Internet Development in China.” 28. See http://www.itu.int/en/mediacentre/Pages/2016-PR48.aspx. https://cnnic.com.cn/IDR/ReportDownloads/. 29. See CETIC, “C15—Proporção de Indivíduos Que Nunca Constine, Josh. 2016. “Facebook Has Connected 40m People with Utilizaram Internet, por Motivos Para Nunca Ter Utilizado Internet.org.” Tech Crunch, November 2. https://techcrunch a Internet,” at http://www.cetic.br/tics/usuarios/2015/total .com/2016/11/02/omnipresent/. -brasil/C15/. Cutler, Ben, Spencer Fowers, Jeffrey Kramer, and Eric Peterson. 2017. “Want an Energy-Efficient Data Center? Build It Under- water.” IEEE Spectrum: Technology, Engineering, and Science References News, February 21. http://spectrum.ieee.org/computing Aker, Jenny C. 2010. “Information from Markets Near and Far: /hardware/want-an-energyefficient-data-center-build-it Mobile Phones and Agricultural Markets in Niger.” American -underwater. Economic Journal: Applied Economics 2 (3): 46–59. https:// Dallas, George. 2014. “Making Sense of Internet Platforms: Network www.aeaweb.org/articles?id=10.1257/app.2.3.46. Effects and Two Sided Markets.” https://georgemdallas.wordpress Alliance for Affordable Internet. 2016. “The Impacts of Emerging .com/2014/06/05/making-sense-of-internet-platforms-network Mobile Data Services in Developing Countries.” Research -effects-and-two-sided-markets/. Brief No. 2. http://1e8q3q16vyc81g8l3h3md6q5f5e.wpengine Dutch Data Center Association. 2017. “The Economic Impact of .netdna-cdn.com/wp-content/uploads/2016/05/Measuring Multi-tenant Data Centers in the Netherlands.” https://www ImpactsofMobileDataServices_ResearchBrief2.pdf. .vijfhart.nl/wp-content/uploads/2017/02/report_-_2017_- Anthony, Sebastian. 2012. “$1.5 Billion: The Cost of Cutting _economic_impact_dutch_data_centers.pdf. London-Tokyo Latency by 60ms.” ExtremeTech, March 20. Elahee, K., and S. Jugoo. 2013. “Ocean Thermal Energy for https://www.extremetech.com/extreme/122989-1-5-billion Air-Conditioning: Case Study of a Green Data Center.” -the-cost-of-cutting-london-toyko-latency-by-60ms. Energy Sources, Part A: Recovery, Utilization, and Environ- Barnett Jr., Thomas. 2016. “The Zettabyte Era Officially Begins mental Effects 35 (7): 679–84. doi:10.1080/15567036.2010. (How Much Is That?).” SP360: Service Provider, September 9. 504941. https://blogs.cisco.com/sp/the-zettabyte-era-officially-begins Ericsson. 2016. “Internet of Things to Overtake Mobile Phones -how-much-is-that. by 2018: Ericsson Mobility Report.” Press release, June 1. 30 Information and Communications for Development 2018 https://www.ericsson.com/en/press-releases/2016/6/internet Kastrenakes, Jacob. 2017. “The FCC Just Killed Net Neutral- -of-things-to-overtake-mobile-phones-by-2018-ericsson ity.” The Verge, December 14. https://www.theverge -mobility-report. .com/2017/12/14/16776154/fcc-net-neutrality-vote-results European Parliament. 2016. “Towards a Thriving Data- -rules-repealed. Driven Economy.” http://www.europarl.europa.eu/sides Katz, Raul, and P. Koutroumpis. 2012. “The Economic Impact /getDoc.do?pubRef=-//EP//TEXT+TA+P8-TA-2016 of Telecommunications in Senegal.” Communications & -0089+0+DOC+XML+V0//EN. Strategies 2 (86): 21–42. https://www8.gsb.columbia.edu/citi Facebook: Internet.org. 2015. “One Year In: Internet.org Free /sites/citi/files/files/Senegal_Telecoms_Report_version%20 Basic Services.” Blog, July 27. https://info.internet.org/en/blog finale%5B1%5D.pdf. /2015/07/27/one-year-in-internet-org-free-basic-services/. Kelly, Heather, and Scott McLean 2017. “Your Browser History ———. n.d. “Where We’ve Launched.” https://info.internet.org Is for Sale, Here’s What You Need to Know.” CNNMoney, /en/story/where-weve-launched/. April 6. http://money.cnn.com/2017/04/05/technology/online -privacy-faq/. Fiegerman, Seth. 2018. “More than 20 States Sue to Stop FCC’s Net Neutrality Repeal.” CNN, January 16. http://money.cnn.com Leavitt, Neal. 2010. “Network-Usage Changes Push Internet Traffic /2018/01/16/technology/net-neutrality-lawsuit/index.html. to the Edge.” Computer [IEEE] 43 (10): 13–15. http://iee explore.ieee.org/document/5604156. Flexenclosure. 2017. “Flexenclosure to Deploy Two eCentre Cable Landing Stations for SSCC in Samoa.” Press release, February Luleå University of Technology. 2014. “National Agenda for the 27. http://www.flexenclosure.com/flexenclosure-deploys-two Internet of Things: Summary of the Project IoT Sweden.” -ecentre-cable-landing-stations-for-sscc-in-samoa/. https://iotsverige.se/wp-content/uploads/2014/12/Agendan -eng-29-sep-2014-korr4.pdf. Garrahan, Matthew. 2016. “Advertising: Facebook and Google Build a Duopoly.” Financial Times, June 23. https://www.ft Mbiti, Isaac M., and David N. Weil. 2011. “Mobile Banking: The .com/content/6c6b74a4-3920-11e6-9a05-82a9b15a8ee7. Impact of M-Pesa in Kenya.” Working Paper 17129, National Bureau of Economic Research, Cambridge, MA. https://ssrn Garret, R. Kelly. 2016. “Facebook’s Problem Is More Complicated .com/abstract=1866089. Than Fake News.” Scientific American, November 17. https:// McGoogan, Cara. 2017. “Yahoo Hack Warning: What Happened www.scientificamerican.com/article/facebook-s-problem-is and Should You Be Worried?” The Telegraph, February 16. -more-complicated-than-fake-news/. https://www.telegraph.co.uk/technology/2017/02/16/yahoo Gartner. 2015. “Gartner Says Australian Organizations to Spend -hack-warning-happened-should-worried/. A$2.5 Billion on Data Center Systems in 2015.” Press release, Millicom. 2016. “Tigo Business Paraguay’s Data Center Awarded May 17. http://www.gartner.com/newsroom/id/3054918. ‘Best Modular Implementation’ by Prestigious International Goldstein, Phil. 2013. “Report: Skype Makes Up One-Third of Sector Magazine.” Press release, September 28. http://www All International Phone Traffic.” FierceWireless, February 15. .millicom.com/media/6762386/Data-Center-Award_-Eng https://www.fiercewireless.com/wireless/report-skype-makes .pdf. -up-one-third-all-international-phone-traffic. ———. n.d. “Investing in State-of-the-Art Data Centers.” News The Green Grid. 2016. “The ‘Big Bad Data Police’ Will Call If We feature. http://www.millicom.com/media/millicom-news Don’t Get More Efficient.” News release, September 5. https:// -features/data-centers. www.thegreengrid.org/en/news room/news-releases/green Moss, Sebastian. 2017. “The Kingdom of Bhutan Opens -grid-big-bad-data-police-will-call. First Government Data Center.” Data Centre Dynamics, Greengard, Samuel. 2010. “Cloud Computing and Developing March 27. http://www.datacenterdynamics.com/content Nations.” Communications of the ACM 53 (5): 18–20. https:// -tracks/colo-cloud/the-kingdom-of-bhutan-opens-first-gov cacm.acm.org/magazines/2010/5/87255-cloud-computing ernment-data-center/98052.fullarticle. -and-developing-nations/fulltext. Nugraha, Y., Kautsarina, and A. S. Sastrosubroto. 2015. “Towards Hellemans, Alexander. 2015. “Why IoT Needs 5G.” IEEE Spec- Data Sovereignty in Cyberspace.” Working paper presented trum, May 20. http://spectrum.ieee.org/tech-talk/computing at the Third International Conference on Information /networks/5g-taking-stock. and Communication Technology. https://www.cs.ox.ac.uk Ingram, Matthew. 2017. “Facebook and Google Need to Be Regu- /files/7463/Towards%20Data%20Sovereignity%20in%20 lated, Says British News Industry.” Fortune, March 9. http:// Cyberspace_Nugraha.pdf. fortune.com/2017/03/09/facebook-google-regulated/. National Statistical Office, Thailand. 2016. “The 2016 Household Internet Society. 2017. “A Case Study in Local Content Hosting: Survey on the Use of Information and Communication Technol- Speed, Visits, and Cost of Access.” https://www.internetsociety ogy.” Ministry of Information and Communication Technology. .org/doc/case-study-local-content-rwanda. http://web.nso.go.th/en/survey/ict/ict_house16.htm. Supply: Data Connectivity and Capacity 31 Netcraft. 2017. “February 2017 Web Server Survey.” https://news Dallas, TX: SIL International. https://www.ethnologue.com .netcraft.com/archives/2017/02/27/february-2017-web-server /language/eng. -survey.html. Syntonic. 2017. “Syntonic Expands to Latin America and Lending O’Brien, Sarah Ashley. 2016. “Why Facebook Was Launching a Mobile Carriers.” January 21. https://syntonic.com/2017/01 Satellite into Space.” CNN.com, September 1. http://money.cnn /syntonic-expands-to-latin-america-with-leading-mobile .com/2016/09/01/technology/facebook-satellite-explosion/. -carriers/. ODIHQ (Open Data Institute). 2015. “The Economic Impact of Telecom Regulatory Authority of India. 2016. “TRAI Releases Open Data: What Do We Already Know?” Medium, November 2. the Prohibition of Discriminatory Tariffs for Data Servi- https://medium.com/@ODIHQ/the-economic-impact-of-open ces Regulations.” http://www.trai.gov.in/notifications -data-what-do-we-already-know-1a119c1958a0. /press-release/trai-releases-prohibition-discriminatory OECD (Organisation for Economic Co-operation and Development). -tariffs-data-services. 2017. “Mobile Data Usage per Mobile Broadband Subscription, Thornhill, John. 2018. “Social Media Users of the World Unite.” 2016: Gigabytes per Month.” Figure 3.27 in OECD Digital Financial Times, February 5. https://www.ft.com/content Economy Outlook 2017, 146. Paris: OECD Publishing. https://doi /ea6c3a0c-0843-11e8-9650-9c0ad2d7c5b5. .org/10.1787/9789264276284-graph53-en. 3GPP. 2017. “Industry Support for 3GPP NR Announce- Ofcom. 2016a. “Adults’ Media Use and Attitudes.” https://www ment.” Press release, December 21. http://www.3gpp.org .ofcom.org.uk/__data/assets/pdf_file/0026/80828/2016 /news-events/3gpp-news/1931-industry_pr_5g. -adults-media-use-and-attitudes.pdf. Uptime Institute. 2012. “Data Center Site Infrastructure Tier ———. 2016b. The Communications Market Report. https:// Standard: Topology.” http://www.gpxglobal.net/wp-content www.ofcom.org.uk/__data/assets/pdf_file/0026/95642 /uploads/2012/10/TIERSTANDARD_Topology_120801.pdf. /ICMR-Full.pdf. van Zyl, Gareth. 2014. “MTN Unveils Africa’s First Solar Data ———. 2018. “Ofcom Kicks Off Auction to Improve Mobile Centre Cooling System.” ITWeb Africa, July 10. http://www Broadband and Prepare for 5G.” https://www.ofcom.org .itwebafrica.com/cloud/517-south-africa/233194-mtn .uk/about-ofcom/latest/media/media-releases/2018/start -unveils-africas-first-solar-data-centre-cooling-system. -spectrum-auction. Vector. 2017. “Operational Performance for the 9 Months Ended 31 Peterson, Tim. 2015. “Facebook Users Are Posting 75% More March 2017.” Market release, April 26. https://www.vector.co.nz Videos Than Last Year.” AdAge, January 7. http://adage /news/operational-performance-for-the-9-months-ended-31. .com/article/digital/facebook-users-posting-75-videos Verge, Jason. 2015. “PLDT Building Eighth Philippines Data -year/296482/. Center for $29m.” Data Center Knowledge, April 28. http:// RICTA. 2015. “Growing the Rwandan Internet Content Hosted www.datacenterknowledge.com/archives/2015/04/28/pldt in Rwanda.” https://ricta.org.rw/sites/default/files/resources -building-eighth-philippines-data-center-29m. /1kwebsites_project_finalversion_0.pdf. Williams, Chris, Davide Strusani, David Vincent, and David Kovo. Sandvine. 2015. “Global Internet Phenomena: Latin America & 2016. “The Economic Impact of Next-Generation Mobile Servi- North America.” https://www.sandvine.com/hubfs/downloads ces: How 3G Connections and the Use of Mobile Data Impact /archive/2015-global-internet-phenomena-report-latin-america GDP Growth.” Chapter 1.6 in The Global Information Tech- -and-north-america.pdf. nology Report. Geneva: World Economic Forum. http://www3 .weforum.org/docs/GITR/2013/GITR_Chapter1.6_2013.pdf. Scott, Mark, and Mike Isaacs. 2016. “Facebook Restores Iconic Vietnam War Photo It Censored for Nudity.” New York Wong, Julia Carrie. 2016. “Mark Zuckerberg Accused of Abus- Times, September 9. https://www.nytimes.com/2016/09/10 ing Power after Facebook Deletes ‘Napalm Girl’ Post.” /technology /facebook-vietnam-war-photo-nudity.html The Guardian , September 9. https://www.theguardian ?_r=0. .com/technology/2016/sep/08/facebook-mark-zuckerberg -napalm-girl-photo-vietnam-war. Shehabi, A., S. J. Smith, N. Horner, I. Azevedo, R. Brown, J. Koomey, E. Masanet, D. Sartor, M. Herrlin, and W. Lintner. World Bank. 2017. Doing Business 2017: Equal Opportunity for All. 2016. “United States Data Center Energy Usage Report.” Washington, DC: World Bank. http://www.doingbusiness.org Lawrence Berkeley National Laboratory, Berkeley, California. /reports/global-reports/doing-business-2017. https://pubarchive.lbl.gov/islandora/object/ir%3A1005775 World Bank Group. 2018. Africa’s Pulse, no. 17 (April). https:// /datastream/PDF/view. www.openknowledge.worldbank.org/handle/10986/29667. Simons, Gary F., and Charles D. Fenning, eds. 2018. “English.” Zenith Media. Various years. Advertising Expenditure Forecasts. Ethnologue: Languages of the World. 21st ed. English. London: Zenith Media. 32 Information and Communications for Development 2018 Chapter 3 Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence Introduction Public recognition is growing that AI simultaneously creates unprecedented opportunities for societal benefit and T he data universe is ever expanding, as chapter 2 illus- grave risks to human rights—we are thus at a critical junc- trates. In fact, it is estimated to double in size every ture in the evolution of these technologies. We must seize the two years (EMC Corporation 2014) with some 2.5 opportunity to shape future use as a force for good, ensur- quintillion bytes of information being generated daily (Kapoor ing that the technologies are leveraged in ways that address 2017). Because we increasingly use digital devices to communi- inequalities and avoid widening the digital divide. cate, buy and sell goods, transfer money, search for informa- tion on the internet, and share our lives on social networks, we The big data revolution leave digital trails or “digital exhaust.” A growing amount of digital data is thus being generated as a by-product of our daily As chapter 1 notes, the concept of big data typically describes lives, but also through the increasing digitization of content data sets so large, or so complex, that traditional data- and the spread of the Internet of Things. This growing volume processing techniques often prove inadequate. The term of data is driving the development of big data analytics and “big data” thus captures not only the large volumes of data artificial intelligence (AI), the subjects of this chapter. now available, but also the accompanying processes and The chapter describes opportunities for harnessing the technologies for collecting, storing, and analyzing it. In other value of big data and AI for social good, and how new words, “big data” is not just about data—“no matter how families of AI algorithms now make it possible to obtain big or different it is considered to be”—it is primarily about actionable insights automatically and at scale. Beyond inter- “the analytics, the tools and methods that are used to yield net business or commercial applications, multiple examples insights,” including the frameworks, standards, and stake- already exist of how big data and AI can help us achieve our holders involved in the field and ultimately the knowledge shared development objectives, such as the 2030 Agenda for generated (Maaroof 2015). Sustainable Development1 and the Sustainable Development Although businesses increasingly are mining the digital Goals (SDGs).2 But ethical frameworks need to be developed trails we leave behind to predict consumer behavior, track in line with increased uptake of these new technologies— emerging trends in the market, and monitor operations any discussion of ethics is not limited to the privacy of the in real time to improve sales and profit margins, big data data, but also relates to the impact and consequences of analytics also holds enormous potential to help understand using data and algorithms—or failing to use them. and address pressing socioeconomic and environmental issues. 33 Big data can help inform policy and interventions that set leverage technologies such as visual, speech, and text recog- us on a more sustainable development path and improve nition, as well as robotics. responses to humanitarian emergencies. Machine learning is one such subdiscipline. Whereas Innovation labs across academia, government, the inter- hand-coded software programs typically contain specific national development community, civil society, and the instructions on how to complete a task, machine learning private sector have been using big data and AI to develop a allows a computer system to recognize patterns and make wide range of applications, from mapping discrimination predictions. Deep learning, a subset of machine learning, against refugees in Europe (UN Global Pulse 2017a) to goes one step further—with deep artificial neural networks, facilitating the rescue of migrants at sea based on shipping based on complex algorithms, computers can learn from data (IOM 2017), detecting fires in the Indonesian rainforest large volumes of data while reaching new levels of accuracy (UN Global Pulse 2016), predicting food insecurity due to (Touger 2018). changing food prices via Twitter (UN Global Pulse 2014), In sum, AI is enabling computer systems to collect, or fighting the effects of climate change.3 Box 3.1 describes analyze, and process large amounts of data in real time to how big data is also being used to predict and respond to recognize patterns, make decisions, and, more significantly, disease outbreaks. to learn from said data and from their own experiences. Meanwhile, recent advances in sensors and imaging technologies and data storage, processing, and transfer tech- The evolution of artificial nologies, as well as complex and self-improving algorithms, intelligence to name but a few, are the range of expanding AI applica- Historically, the term “artificial intelligence” has been tions available today. AI is already incorporated in several applied where computer systems imitate thinking or behav- online products, including Google search, Google Translate, ior that people associate with human intelligence, such as and Facebook’s automatic photo-tagging and translation learning, problem solving, and decision-making. Modern applications. Financial companies rely on AI to produce the AI comprises a rich set of subdisciplines and methods that financial modeling that underpins their insurance, banking, Box 3.1 Using big data to predict dengue fever outbreaks in Pakistan Dengue fever is the most rapidly spreading mosquito-borne viral disease in the world. It is endemic in Pakistan, where human mobility and hospitable conditions for mosquitoes have helped it spread. Those infected typically suffer from severe illness, and mortality rates are high. A partnership involving Telenor Research, the Harvard T.H. Chan School of Public Health, Oxford University, the U.S. Centers for Disease Control and Prevention, and the University of Peshawar used big data to anticipate and track the spread of dengue in Pakistan. The partnership leveraged anonymized call data records from 40 million Telenor Pakistan mobile subscribers during the 2013 outbreak to map the geographic spread and the epidemiological timeline of the disease. The analysis combined transmission suitability maps with estimates of seasonal dengue virus importation to generate detailed and dynamic risk maps, helping to inform national containment and epidemic preparedness in Pakistan and beyond. More broadly, the project illustrates the potential of mobile data to reveal mobility patterns that can help accurately predict the spread of disease. The insights it generated helped predict the spread days or even weeks earlier than traditional means. Source: Adapted from Wesolowski et al. 2015. 34 Information and Communications for Development 2018 and asset management products. Moreover, leading research represent a small but significant innovation in learning hospitals have started using AI tools to help medical profes- about the world around us. Taken together, they provide new sionals diagnose and choose the best course of treatment for ways to detect and respond to world events, influence policy their patients. debates, and drive development, in a way that is both safe Although the current application of AI is mostly limited and fair (figure 3.1, table 3.1). to internet business, digital marketing, gaming, and self- The following sections examine the benefits and applica- driving cars, a wealth of opportunities exist for AI methods tions of big data and AI—including for (a) speech and audio to perform different tasks that can accelerate achievement processing, (b) image recognition and geospatial analysis, of the SDGs and inform humanitarian practice. Box 3.2 and (c) text analysis. They also describe how AI is being describes how AI can help transform traditional sectors, leveraged to support the SDGs and address the emerging such as transport. challenges and risks that accompany the uptake of these technologies. Using big data and AI as a force for Speech and audio processing social good Arguably, one major achievement of big data and AI AI and big data are generating new tools and applica- has been to facilitate real-time translation of a growing tions creating actionable insights, real-time awareness, and number of the world’s languages. Although language predictive analysis on numerous topics for sustainable translation is not an SDG per se, greater language and development and humanitarian action. More and more cultural understanding could help increase the effi- compelling examples illustrate the value of this technol- ciency and effectiveness of development efforts across all ogy to improve early warning systems and inform policy SDGs—for example, by helping to map public opinion and programmatic response. These individual use cases (see box 3.3). Google and Microsoft systems, for example, Box 3.2 Artificial intelligence and the transport sector The proliferation of big data is helping to transform the transport sector. Fueled by data and connectivity, a variety of intelligent transport systems have been introduced as the sector rapidly evolves. Alongside other disruptive technologies, such as connected vehicles and automated driving, these intelligent systems are soon expected to completely change the way people and goods are moved. Big data can be combined with predictive analytics, for example, to optimize cargo transport networks based on projected shipping demand.a Data exchanged among vehicles and infrastructure will soon be used to automatically optimize vehicle routes and speeds in real time, reducing congestion and emissions. In the Philippines, for example, real-time traffic data shared using open source tools is being used to optimize traffic flows in Manila and Cebu City.b In Indonesia, location information from GPS-stamped tweets is being used to reveal commuting statistics in the Greater Jakarta area.c The potential for data-driven intelligent transport systems to transform the world’s transpor- tation systems is immense, particularly if the data is combined with new ways to link disparate data sets and creative methods to visualize data. a. https://www.economist.com/news/briefing/21741139-will-be-bad-news-some-global-logistics-business-going-be -transformed. b. http://www.worldbank.org/en/news/press-release/2016/04/05/philippines-real-time-data-can-improve-traffic -management-in. c. https://www.unglobalpulse.org/projects/improving-transport-planning-with-data-analytics. Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 35 Figure 3.1 The Sustainable Development Goals Source: United Nations. Table 3.1 Examples of artificial intelligence applications for the Sustainable Development Goals Value of artificial SDGs intelligence Case study Risks and challenges SDG 1: Artificial intelligence Combining satellite imagery and machine learning There is a risk of omitting No poverty (AI) can be used to predict poverty in Nigeria, Tanzania, Uganda, segments of the population that to monitor income Malawi, and Rwanda cannot be captured by remote and track policies to Jean et al. (n.d.) combined nighttime maps with high- sensing signatures because of identify progress and resolution daytime satellite images to obtain estimates their lack of footprint or the given successful practices. of household consumption and assets. Using survey sociocultural context. and satellite data from five African countries—Malawi, Nigeria, Rwanda, Tanzania, and Uganda—the study showed how a convolutional neural network can be trained to identify image features that can explain up to 75 percent of the variation in local-level economic outcomes. SDG 2: AI can be used to Detecting patterns in big data saves Colombian rice Overexploitation, based on Zero hunger maximize yields and farmers huge losses local optimization, could lead improve agricultural A project run by the International Center for Tropical to exhausted lands and lack of practices based on Agriculture mined 10 years of weather and crop data resources at the systemic level. multiple data sources. to understand how climatic variation affects rice yields. The project fed the patterns into a computer model and predicted a drought in the region of Córdoba. The center subsequently advised the Rice Producers Federation of Colombia (FEDEARROZ) against planting in the first of two annual growing seasons. This advice saved farmers from incurring significant losses (Palmer 2014). SDG 3: AI can be used to Revolutionizing personalized medicine using AI Overpersonalized medicine could Good health support diagnosis and lead to abuse from the insurance Watson, IBM’s “cognitive computing” platform and well-being personalized medical industry and other stakeholders uses natural language processing to efficiently and treatment. quickly sort through millions of journal articles, based on private personal government listings of clinical trials, and other information. existing data sources to help diagnose patients and provide personalized treatment plans. University of Tokyo doctors reported that the artificial intelligence diagnosed a 60-year-old woman’s rare form of leukemia that had been incorrectly identified months earlier in less than 10 minutes (IBM 2018). (continued next page) 36 Information and Communications for Development 2018 Table 3.1 Examples of artificial intelligence applications for the Sustainable Development Goals (continued) Value of artificial SDGs intelligence Case study Risks and challenges SDG 4: AI can be used to Detecting dyslexia in children in Spain There is the danger that harmful Quality tailor the delivery of media can be easily accessed by Ten percent of the population has dyslexia, a education education based on children. For example, the use of neurological learning disability that affects reading each student’s needs and writing but does not affect general intelligence. YouTube Kids videos optimized and capabilities. Children with dyslexia can learn coping strategies with AI and bots that create to deal with its negative effects. Unfortunately, in long, repetitive, and sometimes most cases dyslexia is detected too late for effective frightening videos meant to keep intervention. Change Dyslexia is a project that uses children entertained for as long cutting edge scientifically based computer games, as possible (Robertson 2017). such as Dytective Test and DytectiveU, that screen and support dyslexia at large scale (Change Dyslexia 2018). SDG 5: AI can help correct for Mapping indicators of female welfare at high spatial AI applications are at risk of Gender equality gender bias in insights resolution in Kenya, Nigeria, Tanzania, Bangladesh, reinforcing existing gender biases derived from big data and Haiti present in the data used to train and nontraditional data A project by Flowminder and WorldPop used the algorithms. sources. geo-located cluster data from the Demographic and Health surveys on rates of literacy, stunting, and use of modern contraception methods to produce high- resolution spatial gender-disaggregated maps, using predictive modeling techniques. The study focused on three countries in Sub-Saharan Africa (Kenya, Nigeria, and Tanzania), one country in South Asia (Bangladesh), and one country from the Western hemisphere (Haiti) (Bosco et al. 2017). SDG 6: AI can predict Monitoring coastal water quality in real time in AI (or simple malware) can be Clean water consumption patterns Singapore used to attack or disable critical and sanitation from sensor data to public infrastructure by means of Project Neptune is a real-time monitoring and prediction optimize water and system strategically deployed around Singapore’s remote warfare. sanitation provision. coastline. The system integrates hydrodynamic and water quality modeling into a forecasting framework that forms the backbone of a central operational management system. Eight specially outfitted buoys act as miniature labs, collecting data on pollutants, including oil and nutrients, and send live updates to the authorities on how they could spread (NUSDeltares 2014). SDG 7: AI can be used Preventing power supply failures in domestic As noted above, critical network Affordable and to make existing railway networks in India infrastructures may be subject to clean energy infrastructure more cybersecurity threats. Aiming to reduce the risk of signal failure, Indian intelligent and energy Railways has trialed remote condition monitoring of the efficient. power supply systems, leveraging AI to predict possible outages. The measure is set to be rolled out on two sections of the Western and South-Western railway network (Economic Times 2017). SDG 8: AI can be used to Optimizing online job searches If algorithms learn hiring practices Decent work optimize recruitment based on biased data that LinkedIn, a well-known business- and employment- and economic for both employers prefers, for example, Caucasian oriented social networking service, uses AI and big growth and jobseekers. names rather than others, it can data to help recruiters automate much of the candidate screening process. The tool is also integrated in make biased hiring decisions. different applicant tracking systems and, for example, automatically synchronizes with the different open jobs, ranking candidates against them (LinkedIn 2017). (continued next page) Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 37 Table 3.1 Examples of artificial intelligence applications for the Sustainable Development Goals (continued) Value of artificial SDGs intelligence Case study Risks and challenges SDG 9: AI can be used Speeding up toy production in Denmark AI will transform and could Industry, to automate and eliminate some jobs. McKinsey A factory in Denmark uses autonomous robots and innovation and eliminate rote or estimates that some 60 percent precision machines to make 36,000 Lego pieces per infrastructure routine work, freeing of all jobs will see a third of their minute, or 2.16 million pieces every hour.a up labor to focus on activities automated (McKinsey more creative tasks. Global Institute 2017). SDG 10: AI can support Accelerating development in Uganda with speech Advances in robotics and AI Reduced translation of less- recognition technology could increase inequality within inequalities known languages societies, further entrenching the UN Global Pulse and the Stellenbosch University to ensure all voices in South Africa used machine learning to develop divide between rich and poor. are accounted for speech-to-text technology to filter the content of public in decision-making radio broadcasts for less-known languages spoken in processes. Uganda. Once converted into text, the information can reveal sentiment around topics relevant for sustainable development (UN Global Pulse 2017b). SDG 11: AI can measure Inferring commuting statistics in Indonesia AI may lead to cascading failures Sustainable traffic in real time, with Twitter of interconnected systems in cities and monitor commuting smart cities. Failures in machine Some estimates for the Greater Jakarta area put the communities statistics, or improve learning algorithms need to population at more than 30 million. In response to transportation the needs of the authorities, UN Global Pulse—Pulse be accommodated in urban services. Lab Jakarta initiated a project to test whether location emergency planning.b information from social media on mobile devices could reveal commuting patterns in the area. The results of the research confirmed that geo-located tweets have the potential to fill current information gaps in official commuting statistics (UN Global Pulse 2017c). SDG 12: AI can improve Supporting smart recycling in the United States AI can also be used to increase Responsible efficiency of recycling with dumpster diving robots the scale of extractive or consumption processes, which can manufacturing industries, creating Spider-like robotic arms, guided by cameras and and production eliminate waste and a larger environmental footprint artificial intelligence, are helping to make municipal improve yields. recycling facilities run more efficiently in the United over time. States. Through deep learning technology, robotic sorters use a vision system to see the material, AI to think and identify each item, and a robotic arm to pick up specific items. The technology could help make recycling systems more effective and profitable (O’Conner 2017). SDG 13: AI and climate science Predicting road flooding for climate mitigation in Heavy computation required to Climate action can help researchers Senegal power AI may lead to increased identify previously Using data from mobile operator Orange, a team energy costs (Lee 2017). unknown atmospheric from the Georgia Institute of Technology developed a processes and rank framework to improve the resilience of road networks climate models. in Senegal to flooding, including recommendations on how to prioritize road improvements given a limited budget. The results showed how roads are being used, how they are damaged, and how policy makers can allocate budget in the most efficient way to repair them (Data for Climate Action 2017). (continued next page) 38 Information and Communications for Development 2018 Table 3.1 Examples of artificial intelligence applications for the Sustainable Development Goals (continued) Value of artificial SDGs intelligence Case study Risks and challenges SDG 14: AI can help detect, Supporting sustainable legal fishing in Indonesia The data collected might be Life below track, and predict the incomplete, as some vessels may Indonesia and Global Fishing Watch—a partnership water movement patterns be undetectable when switching between Google, Oceana, and SkyTruth—are of vessels engaged in cooperating to deliver a vessel monitoring system off their transmitters. illegal fishing. for all Indonesian-flagged fishing vessels and generate data that is publicly available. The project aims to promote transparency in the fishing industry (Global Fishing Watch 2018). SDG 15: AI can be used to map Identifying, counting, and describing wild animals Monitoring technologies can be Life on land and protect wildlife on in camera-trap images in Tanzania used by poachers just as easily land using computer The University of Minnesota Lion Project deployed as conservationists. vision systems. 225 camera traps, across 1,125 square kilometers, in Serengeti National Park to evaluate spatial and temporal dynamics. The cameras accumulated some 99,241 camera-trap days, producing 1.2 million pictures between 2010 and 2013. Members of the general public classified these images via a citizen-science website. The project then applied an algorithm to aggregate the classifications to investigate multi- species dynamics in the local ecosystem (Swanson et al. 2015). SDG 16: AI can reduce Turning information into knowledge and action in Citizen monitoring could be Peace, justice discrimination and Estonia misused to repress political and strong corruption and drive practices (such as voting, In Estonia, government services—legislation, voting, institutions broad access to demonstrations). education, justice, health care, banking, taxes, e-government. policing, and so on—have been digitally linked across one platform, “wiring up” the nation. Estonia is also exploring ways to leverage AI to improve e-government and other public services (e-Estonia 2017). SDG 17: AI should be a public Leveraging partnerships to improve AI for global Collaboration must also result in Partnerships for good. good action. the Goals Multisectoral collaboration is essential for the safe, ethical, and beneficial development of AI. The Partnership on AIc represents a collection of companies and nonprofits that have committed to sharing best practices and communicating openly about the benefits and risks of AI research. Another example is the annual “AI for Good Global Summit”d organized by the International Telecommunication Union, the UN’s specialized agency for information and communication technologies. Note: AI = artificial intelligence; SDG = Sustainable Development Goal. a. https://www.youtube.com/watch?v=whv-krWnq0g. b. For example, mapping apps were reportedly directing people fleeing the California wildfires in Los Angeles, in 2017, to areas that were exposed (Price 2017). c. https://www.partnershiponai.org/partners/. d. https://www.itu.int/en/ITU-T/AI/Pages/201706-default.aspx. are now able to translate over a hundred languages Early models of machine translation used statistical (Li 2016). Also, new systems have been developed that methods that translated words based on a short sequence, perform real-time translations—such as a Skype system that is, within the context of several words before and after that can translate voice calls into 10 different languages in the target word, which did not always work for long and real time (Caughill 2017). complex sentences.4 New neural network architectures, Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 39 Box 3.3 Using machine learning to analyze radio broadcasts in Uganda Radio remains a primary source of information for communities in many parts of the world, particularly in remote rural areas where coverage and access to other forms of connectivity is limited. Radio is also an accessible medium for the millions who remain illiterate. In Uganda, where a majority of the population lives in rural areas, radio is a vibrant platform for community discussion, information sharing, and news broadcasting. Radio talk shows and dial-in discussions are popular forums for voicing local needs, concerns, and opinions. UN Global Pulse collaborated with Stellenbosch University in South Africa to develop speech-recognition technology to automatically convert these radio broadcasts into text for several of the languages spoken in Uganda, including English, Luganda, Acholi, Lugbara, and Rutooro. “Radio mining” consisted of two automated software stages and two human analysis stages. This semi-automated approach allowed a relatively small team of analysts to process many audio recordings quickly and affordably. Several projects were piloted with UN partners to understand the value of talk radio to provide information on topics relevant to the Sustainable Development Goals, such as health care service delivery, response to disease outbreaks, and the efficiency of public awareness- raising radio campaigns, among others. Source: Adapted from UN Global Pulse 2017b. such as long short-term memory, have drastically improved build dynamic population maps and estimate cross-border efficiency. Such systems can now learn from millions of flows of migrants to enable development actors to track the examples and are able to translate whole sentences at a time, spread of disease. This technique was leveraged in southern rather than word by word (Turovsky 2016). Africa to map the movements of cross-border communities to better understand malaria infections patterns (Rango and Computer vision, image analysis, and geospatial data Vespe 2017). Accurate population information is critical for authorities In the environmental field—SDGs 12, 13, 14, and to plan and deliver quality public services and coordi- 15—AI-assisted analysis of satellite imagery can be used nate crisis-relief efforts. However, collecting related data to monitor damage to coastal areas due to floods or traditionally is a long-standing challenge for development typhoons, or drought-affected areas, or the retreat of practitioners and policy makers. For example, gathering wetlands or encroaching land use in deltas or river basins. national household survey data on poverty is typically Combined with meteorological models and large data time-consuming and expensive, requiring elaborate data sets on changes in ocean temperature and currents, such collection and analysis techniques. This exercise is particu- mapping can help improve forecasting and early warning larly challenging in fragile states, where limited capacity systems of future major weather events. Moreover, GPS and security concerns typically hinder data collection and data has been used to analyze traffic patterns to reduce processing. In this setting, for example, satellite imagery has pollution (see box 3.6). been used to gain an overview of population density and Another AI application getting considerable attention assess poverty and access to energy—covered by SDG 1 and is automated or self-driving cars—a potential solution for SDG 7 (see boxes 3.4 and 3.5). optimizing transportation in ways that can minimize car In the health sector—covered by SDG 3—current advan- accidents. Debate is ongoing about what a fully automated ces in medical imaging and computer analysis of tumors can car really is, but considerable progress has been made toward complement and refine radiologists’ analysis. Mobile phone solving problems of visual recognition, object identification, call records have also been combined with satellite data to and reaction processing, which are critical to this endeavor. 40 Information and Communications for Development 2018 Box 3.4 Estimating population counts and poverty in Afghanistan and Sudan In Afghanistan, the United Nations Population Fund and the UN Country Team collaborated with Flowminder, an organization that collects, aggregates, and analyzes anonymous mobile, satellite, and household survey data to generate population maps. The project used survey data, geographic information systems, and satellite imagery data to estimate populations in areas with no such data. In Sudan, the United Nations Development Programme used satellite data to estimate poverty by studying changing nighttime energy consumption. The team used data pulled from nighttime satellite imagery, analyzing illumination values over two years, in conjunction with elec- tric power consumption data from the national electricity authority. The study was also informed by desk research, including similar World Bank work in Kenya and Rwanda. Electricity consump- tion was used as a proxy indicator for income, as poorer households were assumed to be lower energy consumers. The exercise demonstrated how satellite imagery can help measure poverty. Sources: Rango and Vespe 2017; UNDP and UN Global Pulse 2016. Box 3.5 Mapping energy access in India Satellite night-light data has also been leveraged in India. A team from the University of Mich- igan, the U.S. National Oceanic and Atmospheric Administration, and the World Bank Group’s Energy and Extractives Global Practice analyzed the daily light signatures of more than 600,000 villages from 1993 to 2013 (see map B3.5.1). Map B3.5.1 Night lights in India Source: World Bank, Energy and Extractives Global Practice, India Lights Platform (http://india.nightlights.io/). (continued next page) Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 41 Box 3.5 (continued) Electrification trends were visualized on NightLights.io, an open-source platform for processing big data in a scalable and systematic way. The platform features an application programming interface that enables technical partners to query light output. And its interactive maps allow users to explore light output trends. Through the project, the research team gained a high-level overview of rural electrification, compared villages and plot trends, and shared data, which can help inform government policy.a Source: Gaba and Sánchez-Andrade Nuño 2016. a. The platform can be freely accessed and explored at http://india.nightlights.io/#/about. Box 3.6 Cleaning Mexico City’s air with big data and climate policy Mexico City’s congestion, among the world’s worst, worsens local air quality. City dwellers are exposed to twice the recommended level of ozone and fine particulate matter (PM2.5), as advised by national standards and according to 2016 data, resulting in some 10,850 annual deaths. A team of researchers from the University of California, Berkeley, and the Instituto Nacional de Ecología y Cambio Climático in Mexico used data from Waze, a GPS navigation software, to evaluate various transport electrification options based on their ability to reduce urban air pollution and emissions—including (a) the electrification of the entire city taxi fleet, (b) the electrification of public transit buses, and (c) the electrification of all light-duty vehicles. The team first measured the number, location, and duration of traffic jams throughout the city, estimating related emissions using the MOVES-Mexico model. The team then used data from Google’s “popular times” function to map urban population movement. Using this information, the team was able to identify the best policy options and optimal locations for electric vehicle charging stations. Source: Data for Climate Action 2017. Building on humble beginnings and minor innovations of the conversation, the “what” or the “buzz” of the (including cruise control, assisted steering, lane assist, auto- conversation, “how” people are feeling, and “why” the matic braking, and “Traffic Jam Assist”), the race toward a conversation is happening. Conversations are categorized fully automated car is now underway (box 3.7). and discussion topics identified. The technology is being leveraged, among other things, to Text mining and text analysis support agricultural development and build food security— Also known as text mining, text analytics is the science covered by SDG 2. Kudu, a mobile auction market appli- of turning unstructured text into structured data. Text cation, is using text analysis algorithms to match farmers analytics is focused on extracting key pieces of infor- looking to sell their produce with suitable market traders. mation from conversations. By understanding the The system allows any farmer or trader to send a message language, the context, and how language is used in by phone. Once matched, compatible buyers and sellers everyday conversations, text analytics uncovers the “who” are notified. Kudu not only limits unnecessary travel and 42 Information and Communications for Development 2018 Box 3.7 Self-driving cars Human error causes about 90 percent of all car accidents. Artificial intelligence (AI) and autono- mous driving might therefore help reduce accidents and save lives. Self-driving cars have to identify, assess, evaluate, and respond to fast-changing circumstances, and predict likely events in real time. A fully automated car has to master vehicle dynamics, control systems, and sensor optimization. For example, detecting pedestrians from images or video is a very specific image-classification problem. Driverless cars require robust data capacity for image processing and recognition. Navigation and mapping data is also essential, with GPS coordinates used extensively. Mercedes, BMW, and Audi purchased the mapping business Here from Nokia for US$2 billion; Here combines “static” mapping data taken from cars with 3D cameras with live information supplied by a network of connected devices, including cars (Bell 2015). In January 2016, Volkswagen partnered with Mobileye, a technology company that develops vision-based advanced driver-assistance systems, to produce its real-time image-processing cameras and mapping service for driverless cars. Ford became the first manufacturer to road test a fully autonomous car in snow on public roads in March 2016 after working with researchers from the University of Michigan to create an algorithm recognizing snow and rain (Ford 2016). Ford has already tested autonomous Fusion cars on public roads in the U.S. states of Arizona, California, and Michigan. Despite these groundbreaking developments, the move toward autonomous driving is not without its problems. Many worry that a car-centric vision detracts from more sustain- able solutions related to public transportation and urban design (covered by Sustainable Development Goal 11). Driverless vehicles are also likely to wipe out millions of jobs, includ- ing taxi drivers, couriers, and truck drivers, something new policies must address urgently. Moreover, legal frameworks will need to keep pace and be redesigned. Although a few countries are moving to issue new legal frameworks for autonomous driving, significant legal gaps remain. dependency on intermediaries, but encourages competition near real-time social media signals can serve as a proxy for by overcoming critical information gaps. The application daily food prices (UN Global Pulse 2014). was developed by the AI Research Group, which is special- Similar techniques are being used to analyze a host ized in the application of AI to problems in the developing of other development issues. For example, the ability to world and operates out of the College of Computing and monitor public sentiment toward policy measures in real Information Sciences at Makerere University in Kampala, time, via social media, can provide critical information on Uganda.5 the impact of policy and how it is playing out in practice, Analysis of text from Twitter feeds has also been used especially for vulnerable groups or households (box 3.8). to track food prices in real time in Indonesia. UN Global Data from social media can also help estimate the number Pulse worked with the Ministry of National Development of expats around the world (box 3.9). Planning and the World Food Programme to “nowcast” As mentioned earlier, conducting household surveys is food prices based on Twitter data. The outcome was a statis- often expensive. New approaches such as monitoring social tical model of daily price indicators for four commodities: media could help address data gaps in developing economies. beef, chicken, onion, and chili. When the modeled prices Moreover, these approaches may capture marginalized or were compared with official food prices, the forecast and migrating communities not always accounted for by traditional actual prices were closely correlated, demonstrating that means such as national censuses (Rango and Vespe 2017). Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 43 Box 3.8 Monitoring public sentiment about policy reforms using social media in El Salvador In April 2011, the government of El Salvador removed a countrywide subsidy on liquid petrol- eum gas, the most common domestic cooking fuel. Instead of subsidizing prices at point of sale, eligible households were given an income transfer. The reform triggered considerable public debate and controversy. UN Global Pulse and the World Bank teamed up to investigate whether social media signals from Twitter could be used to understand public perceptions and social dynamics surrounding the fuel subsidy reform, specifically reactions and concerns about political partisanship, the level of infor- mation reaching communities about the reform, and trust in government commitment to deliver the subsidy. A taxonomy of keywords was developed to filter Twitter for relevant content. Regional experts were consulted to ensure slang words and synonyms were included in the taxonomy. Tweets were then filtered to assess relevance and isolate content originating from El Salvador. The study suggests that social media analysis, using big data and AI, can help inform policy implementation, as the sentiment observed was similar to public opinion measured by house- hold surveys. Source: Adapted from UN Global Pulse 2015. Box 3.9 Shedding light on migration patterns using social media information Data from social media can be used to help estimate migrant populations. For example, stud- ies based on Facebook data yield estimates of approximately 214 million “expats” in the world (people stating that they live in a country other than their self-reported “home country”), close to the 2017 estimated total of 258 million international migrants globally. Among the issues surrounding the use of social media data to estimate migrant populations are the difficulty in defining who an international migrant is, selection bias, and the reliability of self-reported information. But scholars are working on reducing selection bias via model fitting and results are promising. Source: Adapted from Rango and Vespe 2017. From design to responsible use: analyze enormous volumes of data, which in turn Ethical challenges with using big can improve predictions, prevent crimes and help data and AI governments better serve people. But there are also serious challenges, and ethical issues at stake. There Although we are only scratching the surface of what is possible in are real concerns about cyber security, human the new age of big data and AI, and how they can be leveraged for rights and privacy. . . The implications for develop- social good, we also need to grapple with both the unintended ment are enormous. Developing countries can gain risks and malicious use of the same technology. These benefits from the benefits of AI, but they also face the high- and looming risks were aptly articulated by the UN Secre- est risk of being left behind. tary-General at the 2017 “AI for Good Global Summit”: Algorithm-based systems, powered by big data and We face a new frontier, with advances moving AI, increasingly both learn from and autonomously at warp speed. Artificial intelligence can help interact with their environments, as well as one another. 44 Information and Communications for Development 2018 This tends to generate behavioral patterns that cannot to privacy by big data and AI, thereby ensuring compli- always be predicted or explained. Where this evolution ance with privacy requirements, identifying mitigation in AI will ultimately take us is not yet clear. Some raise measures, and effectively classifying the impacts of the risk of autonomous weapons or viruses targeting data and algorithm use. Including issues of ethics and individuals with a particular defective DNA trait as one human rights in any impact assessment, including a frightening scenario. And rising concerns about the privacy impact assessment, could prove more effective malicious use of AI, for instance, for profiling, merits a than developing a separate analysis or ethical review stronger ethical governance and regulatory framework framework. that covers how related methods are developed and For example, UN Global Pulse builds ethical considera- deployed. The risk of unintended consequences of AI tions into its data practices by conducting a “risks, harms, should be accounted for at each stage of innovation, and benefits assessment,” which may help identify antici- beginning with design. pated or actual ethical and human rights issues that may Technologies and algorithms by themselves have no occur during a data innovation project (UN Global Pulse intrinsic morality—however, technology can be used for 2018). The assessment considers the proportionality of good or bad depending on how it is employed. Looking potential benefits compared to risks of harm from data use, at existing technologies, ethical considerations need to as well as risk of harm from the data not being used. If the address questions such as what life-and-death decisions risks outweigh the benefits, the project does not proceed. self-driving cars make. Although privacy norms have been In its “Guide to Personal Data Protection and Privacy,” the long established to protect personal data from misuse World Food Programme also builds ethics into its proced- and ensure individual privacy in the digital world, ethics ures through the application of humanitarian principles and has become an additional tool in AI applications used to risk assessments (WFP 2015). Although ethics may not have protect fundamental human rights and help make deci- clear-cut rules, when assessing the risk of harm along with sions in areas where law has no clear-cut answers. The UN the benefits “any potential risks and harms should not be Special Rapporteur on the right to privacy recommends excessive in relation to the [likely] positive impacts of data formal consultation mechanisms be instituted “including use” (UNDG 2017, 5). ethics committees, with professional, community and Incorporating privacy by design is also crucial for other organizations and citizens to protect against the innovation applications that operate with limited human erosion of rights and identify sound practices” (Cannataci supervision. The rapidly developing nature of AI algo- 2017). A recent example in which ethics and moral obli- rithms can give rise to algorithmic bias and unverified gations of data handling were included in an official UN results. Similar to privacy by design is the concept of document is the “Guidance Note on Big Data for the AI ethics by design, which suggests seven principles, includ- achievement of the 2030 Agenda” adopted by the UN ing recommendations to proactively identify security risks Development Group (UNDG 2017). The note, the first by using tools such as the privacy impact assessment to official document in the UN on big data and privacy, minimize potential harm. In addition, ensuring oversight stresses the importance of ensuring that data ethics is of the entire data innovation process, from design to included as part of standard operating procedures for data use, is vital to securing true incorporation of ethics into governance (box 3.10). AI systems (Cavoukian 2017). Data ethics should be treated holistically using a consistent Moreover, accountability and transparency are critical and inclusive framework that considers a diverse set of ethical principles that must accompany any AI innova- outcomes instead of an ad hoc approach that only accounts tion project (UNDG 2017). “[T]ransparency builds trust for limited applications (Floridi and Taddeo 2016). Such in the system, by providing a simple way for the user to mechanisms include codified data ethics principles or codes understand what the system is doing and why” (IEEE of conduct, ethical impact assessments, ethical training for 2016). To maintain transparency, the Institute of Elec- researchers, and ethical review boards. trical and Electronics Engineers recommends developing Privacy impact assessments, in general, allow develop- new standards that describe measurable, testable levels of ers and organizations to effectively assess the risks posed transparency so systems can be objectively assessed and Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 45 Box 3.10 Data privacy, ethics, and protection: A guidance note on big data for achievement of the 2030 Agenda 1. LAWFUL, LEGITIMATE AND FAIR USE Data should be obtained, collected, analysed or otherwise used through lawful, legitimate and fair means, taking into account the interests of those individuals whose data is being used. 2. PURPOSE SPECIFICATION, USE LIMITATION AND PURPOSE COMPATIBILITY Any data use must be compatible or otherwise relevant, and not excessive in relation to the purposes for which it was obtained. 3. RISK MITIGATION AND RISKS, HARMS AND BENEFITS ASSESSMENT A risks, harms and benefits assessment that accounts for data protection and data privacy as well as ethics of data use should be conducted before a new or substantially changed use of data (including its purpose) is undertaken. 4. SENSITIVE DATA AND SENSITIVE CONTEXTS Stricter standards of data protection should be employed while obtaining, accessing, collecting, analysing or otherwise using data on vulnerable populations and persons at risk, children and young people or any other data used in sensitive contexts. 5. DATA SECURITY Robust technical and organizational safeguards and procedures should be implemented to ensure data management throughout the data lifecycle and prevent any unauthorized use, disclosure or breach of personal data. 6. DATA RETENTION AND DATA MINIMIZATION Data access, analysis or other use should be kept to the minimum amount necessary to fulfill the purpose of data use. 7. DATA QUALITY All data-related activities should be designed, carried out, reported and documented with an adequate level of quality and transparency. 8. OPEN DATA, TRANSPARENCY AND ACCOUNTABILITY Appropriate governance and accountability mechanisms should be established to monitor compliance with relevant law, including privacy laws and the highest standards of confidenti- ality, moral and ethical conduct with regard to data use. 9. DUE DILIGENCE FOR THIRD PARTY COLLABORATORS Third party collaborators engaging in data use should act in compliance with relevant laws, including privacy laws as well as the highest standards of confidentiality and moral and ethical conduct. Source: Adapted from the UN Development Group 2017. See the full version at https://undg.org/wp-content/uploads /2017/11/UNDG_BigData_final_web.pdf. 46 Information and Communications for Development 2018 the level of compliance can be determined. Although it is A way forward: Harnessing big data harder and harder to keep algorithms transparent because and AI to “leave no one behind” of heavily interlinked and layered processes of algorithmic This chapter has detailed a handful of examples of the many programming, the AI ethics by design approach suggests innovative applications of big data and AI being used to that ensuring the transparency and accountability of algo- inform sustainable development and humanitarian work rithms is essential to determining the intended outputs and globally (see table 3.1 in particular), illustrating the value of preventing algorithmic bias. this technology for development actors. The overall data ethics program may also include The pervasive nature of big data and the rapidly evolving recurring data ethics reviews at every critical juncture, capabilities of AI hold tremendous promise for social impact such as review boards. A similar approach already exists and can drive transformation across many domains, ranging in research institutions and is usually referred to as from health, to food security, to jobs, and action on climate. internal review boards. For example, in their published Scope therefore exists to expand use of this technology procedures for ethical standards regarding data collec- beyond current applications, leveraging big data and AI in tion, the United Nations Children’s Fund (UNICEF) new ways that help us achieve the 2030 Agenda. National adheres to mechanisms for review such as internal and and international development actors should prioritize external review boards as well as the basic ethics training operational integration of these digital innovations into for researchers. Any UNICEF project involving surveys, policy and practice. Doing so will allow them to craft more focus groups, case studies, physical procedures, games, agile and responsive programming, to support anticipa- or diet and nutritional studies is subject to ethical review tory approaches to managing risk, and to find new ways to (UNICEF 2015). measure social impact. However, mainstream, scaled adop- A stakeholder-inclusive approach that features “the tion by policy makers and communities themselves still faces proactive inclusion of users” is also desirable. “Their inter- systemic barriers and pervasive inertia. action will increase trust and overall reliability of these Given their broad applicability, big data and AI necessi- systems” (IEEE 2016). “[T]he context of data use” should tate new forms of interinstitutional relationships to lever- also always be considered, thus requiring human interven- age data and computational resources, human talent, and tion, and at times, context-specific expertise—such as the decision-making capacity. The capabilities of a diverse set of presence of a humanitarian expert during a humanitarian stakeholders can enable the integration of data innovation response or of a transportation planning expert in a project into ongoing policy processes rather than one-time policy that looks at transportation policy (UNDG 2017). decisions (Maaroof 2015). Finally, ethical approaches to AI should be human- Moreover, as adoption of big data and AI increases rights-centric, incorporating substantive, procedural, and and the technology evolves, so do the potential risks and remedial rights (McGregor 2017). Just as misuse of AI issues that need to be resolved. Many question the suitable may lead to harm, nonuse of AI may allow preventable application of this technology, including malicious use, harm to occur. Decisions to use or not use applications and highlight the risk of unintended consequences in this of AI can infringe on fundamental rights. As suggested by rapidly evolving field, where policy makers may struggle the UN Special Rapporteur on the right to privacy in his to keep pace. Although both the supply of and demand for recent report to the UN General Assembly, “commitment data are expanding at “warp speed,” the data ecosystem, to one right should not detract from the importance and as we know it, is still embryonic—with many advanced protection of another right. Taking rights in conjunc- potential applications still more theory than practice. As tion wherever possible is healthier than taking rights in new capabilities and data sources are applied for good— opposition to each other” (Cannataci 2016, 6, 10). But whether to create smarter public services, better early undoubtedly, incorporating ethics into every stage of warning systems, or more effective responses to crises— project design and implementation of AI can potentially development actors must pause to consider the potential mitigate harm and maximize positive impact of rapidly for harm that may arise, for example, from inadequate developing new technologies, ensuring they are used for privacy protection. social benefit. Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 47 To date, no standards exist for the anonymization advertising, for example, where many of these capabilities and sharing of insights from big data in priority indus- were incubated, big data and AI continue to demonstrate tries such as financial services, e-commerce, and mobile their ability to concentrate wealth—and data—in the hands telecommunications—although the latter has done work to of the few and widen inequalities. develop such standards. At the same time, as noted, nonuse Just as misuse of AI may lead to harm, nonuse of AI of these new capabilities and data sources represents at least may allow preventable harms to occur. The challenge is that as great a risk of harm to the public as that potentially aris- misuse of these new tools is already rife online and real harm ing from inadequate privacy protections. New frameworks is being done, while the opportunity cost of failure to use are needed that go beyond privacy and ensure accountability them responsibly is mounting. Clearly, although achieve- and responsible use and reuse of data for the public good. ment of the 2030 Agenda and the modernization of humani- Principles such as responsibility, accuracy, auditability, and tarian practices necessitates responsible use of these new fairness should be core concepts that guide the development tools, it urgently requires a new, rights-centric effort by all of algorithms and AI. The “society-in-the-loop” algorithm stakeholders to ensure innovations meet community needs concept, for example, proposes to embed the “general will” and no one is left behind. Undoubtedly, assessing the ethical into an algorithmic social contract in which citizens oversee impact of AI in addition to privacy protection measures can algorithmic decision-making that affects them. mitigate harm, maximize benefit, and lead to use of the new Developing countries may have the most to gain from technologies as a force for good. the use of new data sources and tools. However, without thoughtful application and critical complements they may also stand to lose the most. To reap the societal benefits Notes of AI—including expected improvements to productivity 1. See https://sustainabledevelopment.un.org/post2015 and innovation—countries must have access to the data, /transformingourworld. tools, and human expertise necessary to support their 2. See https://www.un.org/sustainabledevelopment/sustainable application, as well as viable plans to address the likely -development-goals/. displacement of workers. The availability of data is to a 3. See www.dataforclimateaction.org for the Data for Climate large degree a by-product of digitization, an area in which Action challenge, an open innovation challenge to channel developing countries lag far behind. There can be no mass data science and big data from the private sector to fight climate change, organized by UN Global Pulse with the digitization without universal and affordable access to support of Western Digital Corporation and the Skoll Global broadband. According to International Telecommunication Threats Fund. Union (ITU) statistics, some 3.8 billion people, or just over 4. See Microsoft Translator, “What Is Neural Network- half the world’s population, were still lacking access to the Based Translation?” https://microsofttranslator.uservoice.com internet in 2017 (ITU 2017). /knowledgebase/articles/1099027-what-is-neural-network The way forward must be inclusive. For the big data and -based-translation AI revolution to benefit the most vulnerable people, current 5. See https://kudu.ug/about/. AI research roadmaps must increase attention to method- ologies that can work in data-scarce environments, that can be adapted quickly and with few examples—as in crisis References scenarios—and that can work with incomplete or missing Bell, M. 2015. “BMW, Audi, Daimler Buy Nokia’s Mapping Unit: data (such as “one-shot learning”). Need is also urgent for An Autonomous Future Is Nigh.” Car, December 8. www bridging gender inequalities in big data. More effort must .carmagazine.co.uk/car-news/industry-news/mercedes-benz /bmw-audi-and-daimler-purchase-nokias-here-system-an be made to train younger generations, women and men, -autonomous-future-is-nigh/. to ensure gender equality and the inclusiveness of ethnic Bosco, Claudio, Victor Alegana, Tom Bird, Carla Pezzulo, Graeme groups in shaping AI. Hornby, Alessandro Sorichetta, Jessica Steele, Cori Ruktanon- As the field of data science accelerates, countries must chai, Nick Ruktanonchai, Erik Wetter, Linus Bengtsson, create robust big data and AI strategies to prevent growing and Andrew J. Tatem. 2017. Mapping Indicators of Female inequalities in access and use of these technologies. In digital Welfare at High Spatial Resolution. WorldPop Project and 48 Information and Communications for Development 2018 Flowminder Foundation. http://www.data2x.org/wp-content Global Fishing Watch. 2018. “Vessel Monitoring Data: Indonesia.” /uploads/2017/02/Mapping-Indicators-of-Female-Welfare http://globalfishingwatch.org/indonesia-vms. -at-High-Spatial-Resolution.pdf. IBM (International Business Machines Corporation). 2018. “IBM Cannataci, Joseph. 2016. “A More In-Depth Look at Open Data Watson for Genomics Helps Doctors Give Patients New and Big Data.” Annex 11 in Report of the Special Rapporteur on Hope.” https://www.ibm.com/watson/health/oncology-and the Right to Privacy. A/HRC/31/64. Geneva: Office of the High -genomics/genomics/. Commissioner for Human Rights. IEEE (Institute of Electrical and Electronics Engineers). 2016. ———. 2017. “Report of the Special Rapporteur on the Right to “The IEEE Global Initiative on Ethics of Autonomous and Privacy.” Doc. A/72/43103. Office of the High Commissioner Intelligent Systems: Ethically Aligned Design; A Vision for for Human Rights, Geneva. Prioritizing Wellbeing with Artificial Intelligence and Autono- Caughill, P. 2017. “Skype Can Now Translate Your Voice Calls into mous Systems V.1.” The IEEE Global Initiative on Ethics of 10 Different Languages in Real-Time.” Futurism, April 10. Autonomous and Intelligent Systems. http://standards.ieee https://futurism.com/skype-can-now-translate-your-voice .org/develop/indconn/ec/autonomous_systems.html. -calls-into-10-different-languages-in-real-time/. IOM (International Organization for Migration). 2017. Fatal Cavoukian, Ann. 2017. “Ethical Standards in Artificial Intelli- Journeys: Improving Data on Missing Migrants. Part 1, vol. 3. gence: AI Ethics by Design.” Presentation to UN Global Pulse, Geneva: IOM. https://publications.iom.int/system/files/pdf December 13. /fatal_journeys_volume_3_part_1.pdf. Change Dyslexia. 2018. “The Best Tool to Detect and Improve ITU (International Telecommunication Union). 2017. “ICT Facts Dyslexia Related Skills” (in Spanish). https://www.changed and Figures.” https://www.itu.int/en/ITU-D/Statistics/Pages yslexia.org/. /facts/default.aspx. Data for Climate Action. 2017. “Grand Prize Winner: Electro- Jean, Neal, Marshall Burke, Michael W. Xie, Matthew Davis, Mobility: Cleaning Mexico City’s Air with Transformational David B. Lobell, and Stegano Ermon. n.d. “Sustainability and Climate Policies through Big Data Pattern Analysis in Traffic Artificial Intelligence Lab: Combining Satellite Imagery and and Social Mobility.” http://www.dataforclimateaction.org Machine Learning to Predict Poverty.” Stanford University. /meet-the-winners/. http://sustain.stanford.edu/predicting-poverty. E-Estonia. 2017. “Artificial Intelligence Is the Next Step for Kapoor, S. 2017. “Artificial Intelligence: Re-imagining Big Data’s E-Governance in Estonia.” https://e-estonia.com/artificial Applicability.” CXOtoday.com, July 17. www.cxotoday.com/story -intelligence-is-the-next-step-for-e-governance-state-adviser /artificial-intelligence-re-imagining-big-datas-applicability/. -reveals/. Lee, T. B. 2017. Bitcoin’s Insane Energy Consumption, Explained. Economic Times. 2017. “Railways to Use Artificial Intelligence https://arstechnica.com/tech-policy/2017/12/bitcoins-insane for Preventing Signal Failures.” November 22. https:// -energy-consumption-explained/. energ y.economictimes.indiatimes.com/news/power Li, A. 2016. “Google Translate Adds 13 New Languages Bring- /railways-to-use-artificial-intelligence-for-preventing-signal ing Total to 103.” https://9to5google.com/2016/02/17/google -failures/61747929. -translate-13-languages. EMC Corporation. 2014. “Digital Universe Invaded by Sensors.” LinkedIn. 2017. “Brilent, the Intelligent Candidate Recommen- Press release, April 9. https://www.emc.com/about/news dation Engine.” https://www.linkedin.com/pulse/brilent-inte /press/2014/20140409-01.htm. lligent-candidate-recommendation-engine-edouard-murat/. Floridi, Luciano, and Mariarosaria Taddeo. 2016. “What Is Data Maaroof, Abbas. 2015. “Big Data and the 2030 Agenda for Ethics?” Philosophical Transactions of the Royal Society A Sustainable Development.” Report for UN-ESCAP. http:// 374 (2083): 20160360. http://rsta.royalsocietypublishing.org www.unescap.org/sites/default/files/Final%20Draft_%20 /content/roypta/374/2083/20160360.full.pdf. stock-taking%20report_For%20Comment_301115.pdf. Ford. 2016. “From Autonomy to Snowtonomy: How Ford Fusion McGregor, Lorna. 2017. “Ethical Development of AI.” Statement at Hybrid Autonomous Research Vehicle Can Navigate in the ITU AI for Good Summit, June. Winter.” https://media.ford.com/content/fordmedia/fna McKinsey Global Institute. 2017. “Jobs Lost, Jobs Gained: Work- / us /en/news/2016/03/10/how-fusion-hybrid-autonomous force Transitions in a Time of Automation.” https://www -vehicle-can-navigate-in-winter.html. .mckinsey.com/~/media/McKinsey/Global%20Themes Gaba, K. M., and B. Sánchez-Andrade Nuño. 2016. “Eyes in the Sky / Fut u re % 2 0 o f % 2 0 O r g a n i z a t i o n s / Wh a t % 2 0 t h e % 2 0 Help Track Rural Electrification.” World Bank blog, January future%20of%20work%20will%20mean%20for%20jobs%20 21. http://blogs.worldbank.org/energy/eyes-sky-help-track skills%20and%20wages/MGI-Jobs-Lost-Jobs-Gained -rural-electrification. -Report-December-6-2017.ashx. Better Data for Doing Good: Responsible Use of Big Data and Artificial Intelligence 49 NUSDeltares. 2014. “Neptune OMS for Singapore Coastal Quality.” ———. 2016. “Haze Gazer: A Crisis Analysis Tool.” Tool Series, http://www.nusdeltares.org/projects/neptune/. no. 2. http://unglobalpulse.org/sites/default/files/2-pager%20 O’Conner, M. C. 2017. “Dumpster Diving Robots: Using AI for Haze%20Gazer%20-%20Feb%202017_0.pdf. Smart Recycling.” IQ, July 5. https://iq.intel.com/dumpster ———. 2017a. “Social Media and Forced Displacement: Big -diving-robots-using-ai-for-smart-recycling/. Data Analytics and Machine-Learning.” White paper. http:// Palmer, N. 2014. “Cracking Patterns in Big Data Saves Colombian Rice unglobalpulse.org/sites/default/files/White%20Paper%20 Farmers Huge Losses.” Research Program on Climate Change, Social%20Media%203_0.pdf. Agriculture, and Food Security (CCAFS), Montpellier, France. ———. 2017b. “Using Machine Learning to Analyse Radio https://ccafs.cgiar.org/research/annual-report/2014/cracking-pat Content in Uganda.” http://unglobalpulse.org/sites/default terns-in-big-data-saves-colombian-rice-farmers-huge-losses. /files/Radio%20Analysis%20Report_Preview%20%283%29 Price, R. 2017. “Mapping Apps Are Reportedly Directing People .pdf. Fleeing the Southern California Wildfires to Areas That ———. 2017c. “Inferring Commuting Statistics in the Greater Are on Fire.” Business Insider, December 7. http://www Jakarta Area with Twitter.” https://www.unglobalpulse.org . businessinsider.com/la-fires-gps-apps-directing-people /projects/inferring-commuting-statistics-greater-jakarta-area -areas-fire-lapd-2017-12. -twitter. Rango, M., and M. Vespe. 2017. “Big Data and Alternative Data ———. 2018. “Data Innovation Risk Assessment Tool.” Data Sources on Migration: From Case Studies to Policy Support.” Innovation for Development Guide. http://unglobalpulse.org Summary Report, European Commission. https://bluehub /sites/default/files/Privacy%20Assessment%20Tool%20.pdf. .jrc.ec.europa.eu/bigdata4migration/uploads/attachments/cj UNDG (United Nations Development Group). 2017. “Data delbdgo00hnqazv3u7xi6pd-big-data-workshop-draft-summ Privacy, Ethics, and Protection: Guidance Note on Big Data ary-report.pdf. and the Achievement of the 2030 Agenda.” https://undg.org Robertson, A. 2017. “What Makes YouTube’s Surreal Kids’ Videos /wp-content/uploads/2017/11/UNDG_BigData_final_web So Creepy?” The Verge, November 21. https://www.theverge .pdf. .com/culture/2017/11/21/16685874/kids- youtube-video UNDP (United Nations Development Programme) and UN -elsagate-creepiness-psychology. Global Pulse. 2016. A Guide to Data Innovation for Develop- Swanson, Alexandra, Margaret Kosmala, Chris Lintott, Robert ment: From Idea to Proof of Concept. http://www.undp.org Simpson, Arfon Smith, and Craig Packer. 2015. “Snapshot /content/undp/en/home/librarypage/development-impact/a Serengeti, High-Frequency Annotated Camera Trap Images -guide-to-data-innovation-for-development---from-idea-to of 40 Mammalian Species in an African Savanna.” Nature. -proof-.html. Scientific Data. https://www.nature.com/articles/sdata201526. UNICEF. 2015. “Procedure for Ethical Standards in Research, Evalu- Touger, Glenn Evan. 2018. “What’s the Difference between Artifi- ation, Data Collection and Analysis.” CF/PD/DRP/2015-001. cial Learning (AI), Machine Learning, and Deep Learning?” April 1. https://www.unicef.org/supply/files/ATTACHMENT Prowesscorp, August 3. www.prowesscorp.com/whats-the-di _IV-UNICEF_Procedure_for_Ethical_Standards.PDF. fference-between-artificial-intelligence-ai-machine-learning United Nations. 2015. Transforming Our World: The 2030 Agenda -and-deep-learning/. for Sustainable Development. New York: UN Publishing. Turovsky, Barak. 2016. “Found in Translation: More Accurate, https://sustainabledevelopment.un.org/post2015/ trans Fluent Sentences in Google Translate.” Google, November 15. formingourworld. https://blog.google/products/translate/found- translation Wesolowski, A., T. Qureshi, M. F. Boni, P. R. Sundsøy, M. A. -more-accurate-fluent-sentences-google-translate/. Johansson, S. B. Rasheed, K. Engø-Monsen, and C. O. Buckee. UN Global Pulse. 2014. “Nowcasting Food Prices in Indonesia Using 2015. “Impact of Human Mobility on the Emergence of Social Media Signals.” Global Pulse Project Series, no. 1. https:// Dengue Epidemics in Pakistan.“ National Academy of Sciences www.unglobalpulse.org/projects/nowcasting-food-prices. 112 (38): 11887–92. ———. 2015. Using Twitter Data to Analyze Public Sentiment WFP (World Food Programme). 2015. WFP Guide to Personal on Fuel Subsidy Policy Reform in El Salvador. Global Pulse Data Protection and Privacy. Rome: WFP. https://docs.wfp Project Series, no. 13. www.unglobalpulse.org/sites/default/files .org/api/documents/e8d24e70cc11448383495caca154cb97 /UNGP_ProjectSeries_ElSalvador_Fuel_2015_0.pdf. /download/. 50 Information and Communications for Development 2018 Chapter 4 People and Data Introduction using the information processed and to enjoy more conven- H ow can the data revolution expand economic ience when interacting with organizations (for instance, development opportunities for more people? Can easier interchange of data between platforms or service the increasing collection, analysis, and use of providers). The increasing use of such data by organizations, data—often from individuals1 through digital transactions such as businesses and governments, implies the potential or digital records of offline activities—broadly benefit those for faster and better decisions by these entities. This can help individuals and people? And what risks might arise, such as them improve service delivery, reduce costs and prices, or to individual privacy, and how might they be managed? support process or product innovation, all of which would Conversations about data have become very popular: inter- benefit the people that those organizations serve. est over time in “big data,” as indicated by Google searches, For example, better techniques for tracking how and for instance, has grown one-hundred-fold since 2010.2 More where people drive their cars can inform traffic planning and data is being generated—by people and machines—and management (Shu 2018). Data from people’s online activ- captured, processed, and transferred than ever before. Much ities inform advertising decisions that fund the operation of of this is because of the increasing use of digital technologies many widely used internet services that are “free” at the point by people and organizations globally; indeed, even most of use (Richter 2018). And as digital tools proliferate, indi- analog processes have digital components (such as a visit to a viduals are increasingly able to benefit directly from access doctor’s office leading to a digital drug prescription). to more and new types of data and the information derived But while the data revolution can benefit people, this from it. People can take steps to increase their physical activ- chapter proposes that the structure of data markets might ity and improve their health by using digital pedometers, be raising risks and costs to individuals. People bear many which have now become available on many smartphones of the costs and risks of participating in data markets and, apart from watch-like activity trackers (Wang and Gandhi indeed, might not even be aware they are participating. The 2014; Lyons et al. 2014). They can analyze market trends and poorest also face entry barriers, and it is possible that they make more informed choices about the products or servi- might not benefit from their participation even when it is ces they buy, for instance, when buying books (Chevalier possible. and Mayzlin 2006) or purchasing air tickets (Sengupta and The benefits from data include—at the most general Wiggins 2014). And depending on the organizations that level—the ability of data users to make better decisions use that data, people may benefit indirectly by being better 51 able to navigate the organization’s products and services and through digital networks and how negative impacts might through an expanded set of choices or opportunities, based be reduced. on analysis of the preferences exhibited by collating data on The chapter concludes with a discussion of public consumer and web traffic choices. policies that could rebalance the costs and benefits The possible costs to the user of data collection include to ensure fairer distribution among participants and the loss of privacy, of agency, and of control. Such costs can understand how data marketplaces can focus more on undermine people’s trust in the organizations that collect, people. These choices could determine whether data control, and use data. Indeed, at the time of writing, various will help people—especially the poor—find economic controversies had broken out over data leaks compromising opportunity. Few best practices exist as models, and the privacy of personal data and the biases involved in the hence the chapter will leave the reader not with specific use of data to profile individuals; these have underscored policy prescriptions, but with a better sense of the the risks emerging in the new data-rich economy. These dynamics at play. costs are not always apparent or are distributed in biased ways among participants in data markets. This is because of The data market how those markets have been evolving, with some organi- zations gaining significant power in defining how such data Technological change and evolving business models is collected, used, and shared. Other risks are emerging in Personal data is generated through an individual’s actions this era of data: because of barriers that prevent people— (such as making a payment using a credit card), through especially the poor—from participating effectively in data business processes that digitize analog data (such as medical markets and analog limitations to the benefits of the data histories), or through consequent machine response (such revolution. as call data records). Such data is now increasingly coming The chapter considers several aspects of personal data from use of the internet, wireless sensors, and the billions markets, which run on the personally identifiable data that of mobile phones around the world. As the world gets people generate (figure 4.1 reviews the types of personal more connected, more people are leaving a digital trail, data). It looks at how data markets have evolved, highlights wherever they go and whatever they do. the various players in the data market, and then discusses This data, which has become more voluminous and the benefits and costs for participants in data marketplaces granular over time, piqued the interest of various orga- nizations that saw the financial value embedded in it. By Figure 4.1 Types of personal data the early 1990s, personal data such as telephone numbers and email addresses was widely used for marketing (Seller and Gray 1999). Companies crunched data to predict Health Government • Medical history • Identification number and identity how likely people would be to buy a product, and began • Prescriptions and vaccinations • Address using that knowledge to come up with targeted marketing • Fitness tracking • Civil information (birth, marriage, and so on) messages. As more digital data was collected, organizations • Legal records began to use increasingly powerful computing tools to manipulate and apply that data. Marketing companies Web Mobile phone built richer consumer profiles to predict future purchases • Email • Number and preferred network • Browsing and search history • Call data records and manufacturing and services companies to design and • Content (social profiles, posts, • Location data (GPS) model new products (Davenport, Cohen, and Jacobson photos, and so on • Social media contacts • Contacts, followers, friends • Purchasing history 2005; Accenture 2015). Companies are now using such data to develop services Financial Other powered by artificial intelligence (AI), and the bigger the • Accounts • Home information data set, the better the AI (Elgan 2016). These and other • Transactions • Travel • Debts • Vehicle information innovations have greatly increased the value of data and • Investments • Inferred data, created using other its potential for being monetized, or bought and sold as a • Insurance data points product in its own right. Data continues to gain value as 52 Information and Communications for Development 2018 its potential uses increase. Organizations—including busi- cases, these organizations depend on the data as a necessary nesses, governments, and others—can derive value from input for their operating model, as do online social media, data by applying the insights arising from data’s analysis search engine websites, and various information and news to internal cost and revenue optimization, marketing and sites. Using advertising as a source of revenue, they typically advertising, intelligence and surveillance, and automation.3 compensate people—producers of the data—with free or The application of personal data for online advertising has highly subsidized access to their services. Hence, data has skyrocketed, with the internet now surpassing television as financial value to those organizations, either immediately the leading advertising channel. At a forecast US$237 billion when it is sold to other organizations (such as marketing in 2018, digital ads are expected to grow from 44 percent of companies) or through the services that an organization global advertising revenues in 2018 to more than 50 percent offers others (such as a search engine selling advertisements by 2020 (Magna Global 2017). Facebook and Google tied to search terms).7 accounted for 84 percent of digital advertising revenue in In other cases, data is an input into an operating model. 2017 (excluding China).4 In 2016, Facebook’s advertising Health systems or government services are one example. revenues were US$27 billion (up more than 1,300 percent Their processes are traditionally standardized and have relied since 2010), accounting for more than 97 percent of its total in the past on highly abstracted models of user preferences. revenues.5 Google’s advertising revenues—US$79 billion Data is thus an input into these systems and does not have (growing 180 percent since 2010)—accounted for 88 percent an immediate financial value but has informational value. of its total revenue.6 Combined, the advertising revenues This implies that such services are often performed for a fee, of these two online platforms were on par with the gross whether paid immediately or separately (such as through domestic product (GDP) of Morocco. taxes). However, these interactions do generate significant amounts of data, and thought has increasingly gone into Data market actors creating more specialized services and choices based on that Table 4.1, complemented by figure 4.1, identifies the main data (such as in e-government services) or improving the types of actors operating in the markets built around quality of those abstract models to design improved services personal data and the relationships among them. Using (such as better medicines or treatments). Businesses can these categories of participants, it is possible to illustrate a unlock financial value by generating more effective insights simple model of the data market, as shown in figure 4.2. from data to launch a new product, reduce waste or costs, People produce personal data, the “raw material,” which enable better decisions, and boost innovations. they “sell” (traditionally at zero price) to other market One may say that the value of data depends on how and players who then use that data to derive various benefits. for what the data is used and how well it is prepared (cleaned Individuals also provide “free labor” on many of these online and organized). In either case, however, data can find its platforms—by creating content such as posts and reviews way to other parties. The regulation of those data flows and by uploading photos and videos—that data collectors is the responsibility of data protectors, which can include can “scrape”—extracting data from online sources—to infer rules pertaining to privacy and sharing of specific types of personal traits and preferences. This personal data, along data (such as health or financial data), as well as rules about with the data that individuals generate from their activities, electronic transactions. Data could also be held as an “asset” and that might be inferred from their data (such as their by those who collect it directly or via others, and new rule political or culinary preferences), is the main source of data systems have emerged around concepts such as the “right for these organizations. to be forgotten” by such entities (Kelly and Satola 2017). People do not always directly derive value or benefits Hence, data regulators can also protect people by defining from this data (until recently, as discussed later). But people and enforcing rules around the use of their data. have been deriving indirect benefits from their sale of data in In this construction of the generic data market, orga- services or products that data-using organizations provide. nizations have an opportunity to capture the value from the These benefits are discussed in the next section. data people produce, and they can determine how much of On the other side of the market, the “buyers” are the this value returns to those people. As noted, this data has various organizations that collect and use the data. In some significant financial and political value since it contains People and Data 53 Table 4.1 Typology of actors in the personal data market Actor Description Examples Data Personal data is generated by individuals as they fill People generate data anonymously through sensors, security producers in forms (either online or offline where the latter is cameras, and the like. digitized), through sensors (such as fitness trackers Individuals generate data using mobile phones, credit cards, and home monitors), through using applications internet search,a fitness trackers, and so on. In some cases, civil and services on mobile phones and the internet, society organizations can help produce data, especially among through using credit cards, and from being captured poorer communities.b by security cameras and other sensors. Data Companies and governments collect data in • A bank asking for financial and personal information. collectors different ways. Businesses collect personal • An internet service provider recording web sites a user has information from their individual customers visited. (Accenture 2015). Similarly, governments collect data from citizens for a wide range of purposes. • An information services company soliciting personal information for an individual to open an email or social media account. • Citizens providing birth, marriage, and death details to governments for civil registries. Data Obtain personal data from public and private parties Data mining companies, such as Acxiomc or notoriously Cambridge aggregators to combine for resale to businesses. Some add Analytica,d that collect information from sources such as public (brokers) additional value through analytics. records and consumer surveys to provide insights for clients such as banks, car companies, and retailers (Singer 2012). Data users Businesses who purchase data aggregators’ Businesses and governments for law enforcement.f products. Users of the analyzed data can also play Alphabet (Google’s parent company), Amazon, Facebook, and the role of data collectors and aggregators on the others are data collectors, data aggregators, and users of the market.e analyzed data. Advertisers are major users of personal data to better target online ads. Open data Prepare (for example, anonymize) and make National governments, affiliated agencies, or organizations (such as providers relevant personal data open to use and redistribute. civil society groups). Data Address privacy and control of personal data. National data protection authorities through privacy and computer protectors Protect the interests of individuals that have crime legislation. generated that data or its derivatives. Companies offer products that provide data security, stronger data protection, or information about personal data that is collected. Many tips are available for protecting personal data. However, the decision to provide personal data in exchange for use of some services is still up to the user.g a. Madrigal 2012. b. For example, Twaweza’s “Uwezo” learning assessments, deployed in East Africa, have informed public debate on education quality in Kenya and highlighted similar issues in the media and among politicians in Tanzania and Uganda. In India, the Society for Education, Action and Research in Community Health provides health care to rural and tribal people in the impoverished Gadchiroli district. Since 1989, it has captured both community-based and hospital-based data as part of its community-based, participatory approach to health care, which has been replicated by other civil society organizations in South Asia. c. See https://www.acxiom.com/. d. The Guardian, n.d. e. Governments are not typically among such users, often due to legal restrictions, though they can obtain data for law enforcement; they would more likely acquire data from primary data collectors. f. See https://transparencyreport.google.com. g. Deloitte Review 2013. information on behavior and preferences. Where people do providing and under what conditions (or at what cost)? provide such data voluntarily, it is because they expect to Do they understand the value (monetary and otherwise) gain some of those benefits—whether it is access to online of the benefits they receive? Are they able to assign value services or better medical care, or merely the chance to win to the data they provide in a manner that explicitly a competition. differentiates between their perception of value and the Questions then emerge from the perspective of data actual value of the benefits that they already have or producers: are people aware of what data they are could receive? And what might ensure that maximum 54 Information and Communications for Development 2018 Figure 4.2 The personal data market Government, business, and Data protectors organizations Volunteered Individuals Data collectors Data brokers Data users Produce Observed Collect and monetize Aggregate, analyze, and monetize Add value and monetize Unpaid and paid digital services Advertising Direct and indirect benefits benefits are delivered to those who produce the data? Benefits The following section unpacks the benefits and costs that The data revolution has given more people access to informa- accrue to individuals as they participate in these data tion they can use to make better decisions. This is, first, because markets as data producers. people can use data for its informational value—either directly or through the organizations that serve them—exposing them to new information or by creating new services or products, The benefits, costs, and risks both of which help them make or realize better decisions. for people Second, it is because their personal data has financial As noted above, the main benefits from the data revolu- value, implicit or explicit, that allows them to exchange tion arise from the information value that personal data (sell or barter) personal data for services or products they can provide to either individuals or to organizations that might otherwise have had to pay for. Often this includes a serve those individuals and from the financial value that range of sophisticated online tools that allow them to be it has for organizations. Yet costs and risks to individuals better informed about services or reduce transaction costs exist in the era of expanding collection, flow, and use of (including information sources such as search engines or personal data. These include, as noted, privacy, loss of communication tools such as email services). agency or control, and risk of exclusion from benefiting Table 4.2 summarizes the two forms of value and provides from data’s value. This section notes that people are not examples of how they operate in the data market. always aware of the costs or the benefits of their participa- tion in the data marketplace and, even if aware, might be Benefits due to informational value constrained in their ability to improve the tradeoff due to When data is organized and analyzed it creates information, the structure of the market. which can be an essential input in economic decision-making People and Data 55 Table 4.2 Benefits from personal data to individual Data holds Informational value Financial value Information is derived from the data people People produce data that has financial value for some other party and produce, which could inform decision-making. exchange their data for products or services. Effects Direct: Derived when people use their own Direct: Derived when people share their data (knowingly or otherwise) or others’ data to make decisions (such as with organizations in return for services (for example, people provide data exercise data from a wearable activity tracker in return for access to information services or social networks online); or reviews on a shopping portal). those services are financed through the sale of the data or its derivates. Indirect: People’s data goes to organizations Indirect: People provide data that collectors use or sell on to others, (for example, health care companies, generating economic value that could return to individuals through urban planners, financial institutions, news lower prices or income-generating opportunities, or feed into broader organizations) that use it to improve or economic processes, which could also include innovations that benefit subsidize their products. the wider public. Benefits • Better decisions • Access to digital services • Innovative products • Wider economic benefits for data users that could spill over into opportunities for data producers • Improvement in public services and security; it influences resource allocation, choices insurers process claims faster. And when people share their about technologies, and political choices and informs them personal data with many of today’s online services providers, about the markets that they participate in. When farmers those companies can attract advertisers, giving more people have access to market pricing information, they can make access to many sophisticated digital tools, from financial better choices about when and where to sell their produce. planners to cloud-based storage, often free. Similarly, when consumers have better information about Organizations can use personal data for innovation in the supply, quality, and price of goods or services, they can processes, products, and services. These innovations could make better choices about where and when to buy them. lead to economic benefits for people through lower prices When data from weather monitoring systems feeds into and a better match between products and consumer needs complex models and informs governments, businesses, and (McKinsey Global Institute 2011). For example, TrueCar10 individuals about potential inclement weather, it allows each collects and analyzes individual transaction data to provide to take measures to minimize or respond to damage. When an idea of local vehicle-specific prices so that car buyers civilians have better information on the events taking place know what others have paid for the same car. And various around them and the decisions their political representatives companies are using personal data to design more engaging are making, they can make better decisions about where to or useful products and services (World Economic Forum live, how to get around, how to spend leisure time, and how 2013). In health care, data collected from large groups of to vote. And when young people have better information individuals is improving diagnoses and helping to identify about careers and wages, they can make better choices about treatment options (SAS Institute, n.d.; Warren 2016). what they study. Personal data is being used to improve public service The data revolution is giving more people increas- delivery, enhance policy making, strengthen citizen partici- ingly diversified and context-specific information through pation, and enhance security. For instance, New York City is improvements in data collection, processing, analysis, and planning to use data from devices installed in taxis that use distribution, online and offline.8 Thus far, people have GPS, as well as pick-up and drop-off data from ride-sharing typically benefited indirectly from data, as when orga- apps, to improve traffic management, identify roads that nizations that collect, process, and use data to make deci- need to be fixed, and determine where to focus efforts after sions or inferences about people’s demands or interests9 inclement weather (Marshall 2017). Similarly, in Seoul, the then provide new or better information or expand the set capital of the Republic of Korea, the location of mobile calls of opportunities available to individuals. For instance— and text messages is used to optimize night bus routes. continuing from a previous example—this happens when One popular application gaining use around the world governments improve disaster preparedness or response or is the use of locational information from smartphones to 56 Information and Communications for Development 2018 report problems with local services (Gunawan 2017). Not such as the availability of credit ratings allowing access to only does this pinpoint the exact location of annoyances, credit. For example, most of those online provide personal such as uncollected garbage, potholes, or graffiti, it can also information for access to advertiser-sponsored digital appli- help foster citizen engagement. Social media activity can be cations such as search engines, storage, email, and social “scraped” to alert vulnerable populations, such as informing media. And there is significant personal-data-driven adver- Brazilian Facebook users about the Zika disease.11 When tising sponsored content online, such as news, health, and personal data is used in ways that improve welfare, people education sites of importance to individuals. will again be open to sharing it with public agencies and Personal data also supports a vast ecosystem of digital other organizations. companies (see chapter 5), and is beginning to influ- And people now are increasingly able to benefit from ence firms outside the traditional digital sectors as well. such data directly, using a wider range of progressively The growth of such businesses—fueled by data—implies sophisticated tools to process data and derive their own economic growth that in turn will benefit individuals. The conclusions. This includes making personal finance deci- large information technology and services companies that sions12 or modifying health-related behavior (Piwek et al. use and benefit from personal data have created thousands 2016). of direct and indirect jobs, for example, and have created platforms that have led to the creation of other businesses. Benefits due to financial value Not all are positive developments, with opportunities for Personal data has financial value, mainly placed on it by some to generate fake data, for instance (see box 4.1). organizations that use that data to market products and services to their customers. This financial benefit is typically Costs and risks not available to the individuals that produce the data, but Despite the potential benefits of the data revolution, people, those organizations could “pay back” the producers of data and especially the poor, are often subjected to many costs directly by providing them with access to additional digital and risks or even precluded from partaking of the benefits services, or indirectly through wider economic benefits, described above. This stems mainly from how individuals Box 4.1 Income-generating opportunities Some people benefit financially and directly from their ability to earn revenue from the data economy. This includes a handful of services that provide money (or discount coupons) in exchange for personal information. People can set up websites and receive income from personal-data-driven advertising tools such as Google’s AdSense. Freelancers can earn money from jobs in data-related areas on Mechanical Turk (www.mturk.com); Upwork a freelance broker reported that jobs associated with data and artificial intelligence were among the fastest growing in the fourth quarter of 2017 (Upwork 2018). And potential could exist for outsourcing analytics projects; a data scientist in India, for example, reported earning US$200 an hour for overseas jobs (Leber 2013). But, while individ- uals could get a financial return for their own personal data, they might also do so with false data. Income can be made from ethically questionable activities such as using fake accounts or reviews to influence social media. For instance, the #richkidsofinstangram handle was used by social media influencers to attract unwitting users to invest in dubious online trading schemes.a Estimates of fake accounts—also created by governments and criminals—range from almost 50 million for Twitter to about 60 million for Facebook (Confessore et al. 2018). a. https://www.theguardian.com/news/2018/apr/19/wolves-of-instagram-jordan-belmont-social-media-traders. People and Data 57 have been largely dependent on the organizations that collect substandard data on a population, with entire groups of or use their personal data as gatekeepers to realize the bene- people invisible, such as unemployed women, indigenous fits of those data, and to act on their decisions. populations, or slum dwellers. The costs and risks stem from two issues involving these The poor also often face constraints on how they use organizations: first, the limitations in how the analog world data—even if they are aware the data exists. This is because permits people to benefit from the data revolution, and growing data flows often do not reach them, due to weak second, the unequal power relationships between people institutions or constraints on the functioning of markets. and these organizations. The first can be discussed briefly, as For example, if government weather data is not made public its resolution requires a shift beyond the data economy itself; quickly, it will not benefit them. Or if disaster preparedness the focus instead will be on the second. and response systems are not in place or fail to operate because alerts do not reach people quickly, even having that Risks arising in the analog world data will not expand opportunities in a way that would allow A key risk in the data economy is that missing analog most to benefit from them. complements, such as limited literacy, can constrain the extent to which people can realize benefits from digital data Costs and risks arising from the data market(s) markets.13 For instance, if organizations do not function well status quo or are in uncompetitive markets, the collection of more data Costs are embedded in how data is shared and consumed, might not improve flows of information or decision-making because of the structure of the markets in which the data by individuals, nor will it create incentives to deliver the is used. These costs might not be transparently disclosed to expected benefits. In such a market, people may perceive individuals (data producers), or they might have unintended the value of their data as low, because of information asym- consequences for the way that the data market functions. metries, and many may give up their data unknowingly or Several costs can be identified: loss of privacy, loss of control, without expecting an appropriate return. loss of agency. When these costs are disclosed or uncovered The poor also face risk of exclusion: the barriers to (especially unintentionally, such as through data leakage, entry in data markets are often too high for them, as they or deliberately through hacking), they could undermine do not have access to digital technologies or they lack the the functioning of the digital ecosystem supported by data skills to use data and convert it into relevant or useful markets due to a loss of trust in the participants in those information. Although the use of new technologies has markets. exploded across the globe in the past 10 years, the price Concerns about privacy have been central to discus- to access this data is still prohibitive for many. In Bolivia, sions about the data economy. Cases exist in which people Honduras, and Nicaragua, for example, a mobile broadband may provide personal data willingly—to government offi- subscription exceeds 10 percent of average monthly GDP cials, health care workers, or marketers. For example, they per capita, compared with France and Korea, where it is less might trade it, knowingly or unknowingly, for access to than 0.1 percent (see figure 5.6). Many people—especially online information services. Collection of such data allows women, people living in the 40 percent of the population data-driven services to improve. But this may mean people with the lowest incomes, or people with disabilities—lack lose some privacy willingly. And it has also made securing the digital tools or literacy to use technology.14 People who and protecting personal data increasingly important for over-share online data concerning their sexuality, eating and all kinds of organizations, both data collectors and users. drinking habits, or their taste for high-risk sports may be Incidents in 201715 and 2018 have shown that the personal unwittingly excluding themselves from insurance coverage, data of millions of people could be accessed—legally, or at least raising their premiums. accidentally, or illegally—including through means that One consequence is that digitally excluded populations neither the individuals nor data collectors might have increasingly risk exclusion from data sets created from been aware of. Most notable is the use of data collected mining digitally generated information that might be used through personality profile surveys on Facebook for to enhance their livelihoods. And this makes many devel- targeted political advertising campaigns (Caldwelladr and oping countries “data poor” themselves; that is, they have Graham-Harrison 2018). 58 Information and Communications for Development 2018 Much personal data, held and used as it is in financial, individual’s preferences, discouraging experimentation and health, or public services organizations, is sensitive, and reinforcing segmented stereotypes, often hidden from view privacy has therefore been recognized as a fundamental both in what the sources of data are and how the algorithm human right deserving protection.16 Loss of privacy risks itself works. At the time of writing, discussions had grown becoming a negative influence on the behavior of others about how algorithms on some platforms might influence or organizations, such as through exclusion of people from significant choices, such as voting.18 And even if the more access to services, social threats (bullying and stalking), or serious of these claims are ultimately unproven, the working employment hiring or firing decisions.17 of many of these algorithms is not clear (as well as what Transparency about what data is being collected, from biases might inadvertently or purposefully exist). whom, and about how it could be used is critical. However, These hidden costs, when they are disclosed, are often much of the data people generate is now automatically then accompanied by significant negative publicity for the created through their actions and often does not request organizations involved. This could undermine the provision explicit permission for collection or sharing with others of such products or digital services—dependent as they are (beyond accepting terms and conditions, often wordy and on personal data—because people lose trust in those services. complicated). Because digital data is effectively permanent Theft of personal data, its growing accumulation and analysis and can be replicated infinitely, its use can extend far beyond by companies, and the spread of fake information increasingly what was earlier possible with analog records. Such loss targeting specific groups of people lowers trust for governments of control occurs as people give data away unwillingly or that people feel are not doing enough to protect them and for unknowingly, and, hence, lose control over it, are not aware companies they feel are misusing their data (Mineo 2017). of how or when it will be used or by whom, and are unable Underlying many of these risks is the imbalanced struc- to engage in its secondary use. ture of many data markets. Increasingly, private organiza- One example is Meitu, a photo-enhancing app that requests tions are holding and using data, and these organizations are access to far more data than needed, such as GPS location, cell not subject to democratic pressure (as many public institu- carrier information, Wi-Fi connection data, SIM card infor- tions are), and increasingly are subject to winner-takes-all mation, and personal identifiers that could be used to track pressure in network industries. As noted, individuals are people’s devices and sell the data without them realizing it. often unable to negotiate better terms and conditions related Users have control over whether to use an application or not to their data or create better trade-offs between their privacy, as well as to adjust privacy settings within applications, but the control, agency, and access to services. Better informed and configurations can be complicated or unwittingly bypassed. targeted regulation is part of the solution, given the collective Often, individuals are unable to deny an organization control action problem that occurs when large numbers of people of their data, sometimes exclusively, without giving up access engage with such organizations or networks. The next to all of its services; no options are available, especially in the section discusses other protections that might be needed. online world, where terms and conditions to give up control are frequently “take it or leave it.” Remedies Loss of agency happens when algorithms or the input data causes people to lose control over their actions or Vibrant debate is now ensuing about what public policies restrict their ability to determine their own choices. Such could help respond to these failures within and outside the loss is reinforced by the development of algorithms that are data market, and how regulations may be applied in this starting to offer choices to people for everything from what sector, which up until now has been largely unregulated. movies to watch, which news sources are relevant, what to Appropriate policies—helped by emerging technologies— buy (Boffey 2017), or which web pages might offer the infor- could lead the data revolution to expand economic oppor- mation they seek (Naone 2011). tunities for more people. Part of this could be achieved by Those algorithms are developed based on models of making the costs and benefits transparent and redistributing personal preferences, using user data, that are abstractions them more fairly across different players in the market. of individual behavior. Such algorithms may frequently be Specific remedies could help address or minimize the inaccurate, no matter their sophistication. They model an risks and costs to individuals arising from the ways data People and Data 59 markets function today. Areas that a personal data policy Key attributes of such a legal framework include protection could address include overcoming the identified market of personal data collected by organizations, such as effective failures—loss of privacy, control, and agency; exclusion from and appropriate security to protect the data from theft and participation in the market; and unfair distribution of the misuse. It is also generally accepted that organizations need market benefits among data market participants. to keep personal data accurate, relevant, and updated. Data But little consensus exists for now on what remedies will subjects must be able to access and correct their personal work, and some approaches are yet to be tested. And current data. Widely cited frameworks to define the rules around the data policies are highly fragmented, with diverging global, privacy of personal data include the European General Data regional, and national regulatory approaches. Moreover, Protection Regulation (GDPR) (EU 2018); the APEC Privacy these remedies do not directly address the unequal power Framework (APEC 2015); and the OECD’s Privacy Guide- of individual users versus the organizations (global plat- lines (OECD 2013). The Council of Europe’s Convention forms or states), an underlying issue in data markets. This 108 is a foundational data protection initiative, with a treaty issue might only be addressed through strong regulatory or that opened for ratifications in 1981 (COE 1981). The treaty large-scale user action; but, again, little consensus exists on intends to “secure in the territory of each Party for every indi- how these might be achieved.19 Table 4.3 and the rest of this vidual, whatever his nationality or residence, respect for his section outline emerging responses. rights and fundamental freedoms, and in particular his right to privacy, with regard to automatic processing of personal Privacy data relating to him (‘data protection’).”20 Privacy protections have been typically ensured through The GDPR, which came into force in May 2018, enables legal frameworks. A global survey, reported by UNCTAD better control over personal data, entitling individual (2017) shows that data and privacy protection legislation protection of anonymity, pseudonymity,21 and rights to has been put in place in more than 100 economies, 66 devel- request and erase personal data (“right to be forgotten”). oping or transitioning (see map ES.1). More than one-fifth Another novel feature is data portability, giving individuals of economies, primarily developing ones, had no legislation, the right to request that their data be transferred to another and few have developed comprehensive data protection laws. controller and for data controllers to use common formats. Table 4.3 Risks and remedies Risk Remedy Example Loss of privacy Legal frameworks to protect personal data from theft and misuse, to European General Data Protection require consent for collection and use, to keep personal data accurate and Regulation (GDPR) (EU 2018); APEC relevant (where data subjects can access and correct their personal data), Privacy Framework (APEC 2015); OECD to define how such data can flow (including across borders), and to specify Privacy Guidelines (OECD 2013). the mechanisms to assist individuals if violations occur. Loss of agency Informing individuals about when and how data is collected and used, None, although some companies such including how their experiences are modified by algorithms based on that as Google do now allow users to “turn and others’ data. Allowing users to switch off such algorithms or hold back off” personalized search results, for their data from being used. Clarity about data sources to minimize the risk example. of fake data or its derivatives influencing decisions. Loss of control Legal frameworks limit the collection of personal data, and limit use and Canadian Personal Information disclosure to specific purposes. Data subjects should be notified about Protection and Electronics Documents the purpose and disclosure of the data collection and can opt out of data Act; European GDPR. sharing between the data collector and other companies. They can also choose to be forgotten. Loss of trust Reducing personal data breaches, business codes of conduct where Data Science Code of Professional regulation is weak or vague, acting on feedback from user communities. Conduct.a Exclusion Connecting people to the better-quality, affordable internet. Universal technology access programs and digital literacy training. a. See http://www.datascienceassn.org/code-of-conduct.html. 60 Information and Communications for Development 2018 Cross-border personal data flows are also regulated, with New York City trains librarians, in turn, to provide guidance onward transmission generally only permitted if the recipi- on protecting personal data to the largely vulnerable patrons ent country has adequate data protection laws. Businesses that utilize libraries’ internet services.26 Some applications that do not comply with the regulation face significant fines. allow individuals to switch off predictive algorithms. For The right of an individual to privacy is often balanced example, Google allows its users to delete their past searches with the need to secure the greater public good. For example, or prevent saving of searches27 or allows users to turn off even the Council of Europe’s Convention 108 permits personalized search results that might create an “echo cham- restrictions in cases when “overriding interests (e.g. State ber” for users by limiting their exposure to new sources of security, defense, etc.) are at stake.”22 In other cases, privacy information.28 rules permit irreversibly anonymized data to be used for research or public interest activities (EU 2018, Art. 26). This Exclusion balances the interests of individuals in safeguarding their Exclusion of individuals from data markets can be overcome privacy with the benefits of being able to use personal data, in different ways. It is estimated that well over 2 billion people as described in the preceding sections. did not use the internet at all in 2016, either because they Beyond legal frameworks, however, new approaches are had no access, could not afford it, or did not know how or emerging. This helps in areas given institutional capacity want to use it. A significant proportion of these people live in limitations, the difficulty in regulating across borders, rural areas of developing countries, where levels of internet and the “take-it-or-leave-it” nature of many services. For infrastructure and incomes are often low. Exclusion from the example, online services that embed privacy into their data market can be overcome through introduction of infor- designs have emerged in messaging or search.23 A more mation and communication technology, particularly mobile detailed discussion is found in chapter 6. telephony and the internet, among lower-income groups and connection of more people through inexpensive phones. Control Governments need equally to tackle the challenge of To overcome loss of control, collection of personal data people who have the needed infrastructure within reach should be transparent, and use or disclosure limited to but do not use the internet because they lack digital literacy. specific purposes. Individuals should be notified about the This could be done through creation of awareness about purpose and disclosure of the data collection. One example data-driven services (such as social networks, public servi- is Canada’s Personal Information Protection and Electronics ces, search engines), as the Indian government’s Digital Documents Act passed in 2000 (passed by the Privacy India Program of 2015 does. The program helps farmers get Commissioner of Canada). Under the act, individuals have access to information about different wholesale markets in the right to access the information held about them, chal- their community through digital apps on smartphones and lenge its accuracy, and give consent for personal information helped cut out middlemen (see Reuters Market Light 2015). to be collected. Organizations have obligations to ensure data Farmers can use this information to make better choices and security, limit the data they collect, use personal data only for not be beholden to centuries-old systems (Bergvinson 2017). the purposes consented to by the consumer, and not retain By the end of 2015 the program had already helped increase the data when purposes for collection are no longer in effect farmers’ incomes 5–25 percent.29 (Green 2018). The EU’s GDPR also enhances individuals’ control over personal data by enabling the “right to be forgot- Trust—and the dominance of digital platforms ten,” permitting them to control what personal data is avail- During the writing of this report, many episodes under- able online or with data users.24 The rules also allow users to scored the scale of the personal data economy, but also control how personal data is used by those organizations.25 undermined the trust that people have in the organizations that have grown significantly in the data market. These Agency episodes have included massive leaks of personal data, Loss of agency can be averted by educating individuals in discovery of unapproved access to private data, attempts at data collection methods and in how algorithms modify their manipulation of ostensibly neutral information sources, and experiences based on their data. The Data Privacy Project in sharing of personal information. The scale of these episodes People and Data 61 is significant, given the reach and popularity of the organiz- US$410 per annual visit, US$30 per visit for quarterly ations and platforms that they involve, such as Experian or assessments, and US$10 for filling in questionnaires (Lomas Facebook. 2017). Debate about the implications of these episodes is only just Businesses are finding that their customers are becoming beginning, and focusing on privacy of personal data, control more informed about the use of their data and the potential over who accesses and uses people’s data, and the agency of monetary value of it, and expect value in return for data used users. In one account, the organizations involved in transgres- to target marketing and for data sold to third parties (Morey, sions might have been unaware themselves of the potential Forbath, and Schoop 2015; Accenture 2015). Companies for trouble or unable to prevent it. But such accounts do little may also begin to find that they lose customers when they to shore up trust in these services. Even so, the scale of orga- fail to keep data secure; however, the winner-takes-all nature nizations’ networks and their importance might lead people of many of the platforms and services in use today might to continue using them, even if less willingly. mean that an exodus might not occur often or easily. It might be possible to instill greater trust through For individuals, the biggest benefit is regaining control actions to remedy some of these other risks. It might also be of personal data. A second gain could be more accurate possible to seek ways to manage data more collaboratively, data, as individuals would have a greater incentive to keep for instance, adopting a code of conduct (such as the Data it up to date to better monetize it. This protects people in Science Code of Professional Conduct of the Data Science instances in which out-of-date information might be used Association) (Data Privacy Project 2018), and with more against them (such as applying for loans or insurance). More transparency, in how data is managed and used. As the next comprehensive information could also expand the scope section discusses, this may involve moving toward a more of applications and services. Third, personal information balanced personal data market in which users regain control would be centralized and simplified using personal data over their data. management software. Individuals would have fewer pass- words to keep track of.31 Thus, it should be possible for people to act as data- Toward a more balanced data producing entrepreneurs—having a data profile, personal market data management software, and an online wallet—and Emerging trends suggest new opportunities for individuals exchange the data for money, discount coupons, or free appli- to regain control of their personal data, giving people more cations and services. The World Economic Forum (2011) has power as actors in the data market. People are looking for proposed the concept of a data bank account, in which an ways to keep their data secure and to monetize it and to get individual’s data would “reside in an account where it would better value in exchange for the personal information they be controlled, managed, exchanged and accounted for.” provide (Whitler 2016). Newer business models—driven One challenge lies in determining the value of personal by technological advances and people’s greater awareness of data. In Italy, a team of researchers monitored a study group the transactions and value of data markets—are prompt- that auctioned off smartphone data for two months, with ing creation of a more balanced market for personal data. the median bid across all data categories of €2 (US$2.72) However, scope remains for greater coordination or even (MIT Technology Review 2014). One individual sold his aggregation of data streams and sources to maximize value. personal data on a crowdsourcing site for US$2–US$200 Emerging business models allow people to control and (depending on the amount and frequency of the informa- directly sell their data to businesses. Companies such as tion), earning US$2,733 from 213 backers in one month, Datacoup30 enable users to sell their personal data for a or an average of US$12.83 per backer (see Zannier, n.d.). monthly fee, for example, data generated through social Another study uses operating metrics from Experian and media activity and credit card transactions (Simonite 2014). Facebook, companies whose revenues are largely generated Another example of this is Alphabet’s Project Baseline, from personal data, finding that the average revenue per user which collects laboratory results and real-time health data of both was about US$6 a year (Roosendaal, van Lieshout, from individuals wearing a special wristband. Participants and van Veenstra 2014). Another perspective on personal in the study share their health data for two years and receive data valuation is total global digital advertising revenue 62 Information and Communications for Development 2018 (US$178 billion; see Magna Global 2017) divided by the Products also exist for individuals to protect at least some number of internet users around the world (3.4 billion),32 for personal information from internet service providers (Kalia an average of US$53 in 2016. 2017). In addition, cookie controls (Brandom 2017) and ad Personal data does not have a uniform value and varies blockers will allow users to block online marketing gener- according to several variables, such as type of information ated from their personal information (Rosenwald 2015). and income of user. Data from Facebook confirms the latter, Some individuals have consciously decided to restrict with the company having different average revenue per user sharing of personal data. Ironically, many people involved in depending on the region. In the end, the value of personal the social media or technology industries limit their use of data will be determined by what purchasers are willing to these services or systems because of concerns about psych- pay. This will become more apparent with the emergence of ological and other dangers caused by services using their global, regional, and national markets for personal data, in personal data (Lewis 2017). These trends may initially lack which data collectors would review the data available and the scale of the large internet companies, but could grow purchase directly from individuals or third parties they have as more individuals weigh the tradeoff between sacrificing entrusted the data to. Personal data management software personal data for unpaid services. that individuals can operate themselves or where firms act as trusted custodians for users who lack the skills are already Looking to the future on the market (Lehtiniemi 2017). It is certain that large internet companies will resist It is possible that the future would lead to greater democra- individuals’ greater control over personal data. Collecting tization of access and use of personal data. This could lead personal data is at the center of these companies’ business to more data sharing and, eventually, transform individuals models, driven by the willingness of individuals to sacrifice from consumers of data to both consumers and suppliers personal data for unpaid services. Developed and developing of data. As data suppliers, individuals would be able price countries also appear to be split over the threat to businesses the data they produce and share with businesses or govern- of individuals monetizing their personal data. In developed ments. nations, it is less of a threat to businesses, with bigger worries The market power of private organizations, especially in government regulation, cyberattacks, and personal data data collecting platforms and networks, possibly, could also protection applications. But in Brazil, China, and India, be moderated. This is possible through the emergence of individuals charging for their personal data is among the top competition (such as new social networks or online service business concerns. providers), regulation by governments or the platforms However, the unbalanced personal data market could themselves,36 and shifts in consumer preferences en masse, lead to greater disenfranchisement among individuals. This which could privilege privacy or control over access. Any in turn could lead a growing number of individuals to opt such shift will emerge out of negotiation among market out of the existing arrangement. players, but should ideally seek to balance innovation by Tools are already available that give people greater control these firms with respect for individuals’ rights. over their personal information. For example, a Swedish Recent trends are shifting the value distribution to the company claims, in a few clicks, to be able to find and delete producers of data, in terms of such things as better health accounts created using Gmail (see Neal 2016). Stricter poli- care (cancer research, medical treatment, or diagnosis) cies about sharing personal information are available with and better public services (such as traffic and road plan- free email,33 office applications,34 and browsers.35 Scope ning, water planning). Technologies such as micropayments also exists for paid tools with tighter privacy controls, as may also lead to innovation in this area, as noted, possibly users might pay for applications and services that protect allowing people to directly sell their data to businesses and different types of personal information. One study found governments in the future. The rise of AI and the Internet of that individuals in the United States would pay most to Things will help individuals trade personal data and receive protect government identification, those in India for credit personalized services based on personal data. card information, and in Germany and the United Kingdom, Apart from the technological aspects and drivers of this for medical records (Morey, Forbath, and Schoop 2015). change, the personal data market that may emerge would People and Data 63 benefit from more consistent structuring and organizing of to be in place to ensure that people can use the informa- data. This is because such organizations—focusing on the tion generated through the exponentially growing streams benefits to people rather than to organizations alone—could of data. The digital data revolution might be upon us, but help aggregate or combine data across platforms and permit people will also need reform in the analog world to effect portability. real change in their lives. Already, data can be accumulated and cross-referenced across various financial services and platforms to detect oppor- Notes tunities to maximize returns on investment. For instance, 1. The other major source of data is machine generated, which online personal finance tools have begun to link people’s bank, accounts for an increasing share of the global total. This securities, retirement, and credit card accounts to provide includes data that is generated automatically, without human ideas and offer products or services to budget better, increase intervention. access to credit, or identify investment opportunities.37 2. See https://trends.google.com/trends/explore?date=all&q=big%20 But we could go further: linking personal data about data,open%20data,data%20analytics. physical movement collected through phone location or 3. See https://www.mckinsey.com/~/media/McKinsey/Business health tracking, combined with data about transportation %20Functions/McKinsey%20Analytics/Our%20Insights use, could be combined to provide people looking to exercise /The%20age%20of%20analytics%20Competing%20in%20 with ideas about adopting a routine that increases walking. a%20data%20driven%20world/MGI-The-Age-of-Analytics -Full-report.ashx. Shopping patterns across various stores could be combined to provide better choices or insight to people about ways to 4. See “Ad Spend Forecast Update 2018: DOOH, Google and Face- book Drive Growth” at http://www.jcdecaux.com/blog/ad-spend save by changing the locations or timing of their purchases. -forecast-update-2018-dooh-google-and-facebook-drive-growth. The potential opportunities to merge data sources and 5. See https://s21.q4cdn.com/399680738/files/doc_presentations improve decision-making holds promise, again, with the /FB-Q4’16-Earnings-Slides.pdf, slide 8. caveat that the costs and risks need to be managed. 6. See https://abc.xyz/investor/pdf/2016_google_annual_report .pdf, page 22. People as a focus for data markets 7. For one example, see Google’s AdWords service at https://en The data revolution holds great promise. When better data .wikipedia.org/wiki/AdWords. is available to people, they can make better decisions and 8. For example, financial transactions are being increasingly find the information needed to improve their economic and digitized—through mobile money and credit cards—or social lives. The technological tools to realize these benefits sensors in buildings, roads, or wearable devices are collecting exist today and will develop further. As more people connect massive troves of increasingly high-resolution data about our to the internet and new ways of collecting, managing, movements or activities in the offline world. and analyzing personal data become commonplace, more 9. For example, Dewey (2016). people, including the poor, will participate in the growing 10. For information, see https://www.truecar.com/. data economy. 11. See “A Case Study: Data and Social Media Can Lead to But these changes will not come without their risks and Healthier Lives” at http://neo-assets.s3.amazonaws.com/news costs. Without measures in place to protect privacy, agency, /FB-UNICEF-Big.png. and control over data, the risk is that businesses and orga- 12. See https://www.investopedia.com/personal-finance/personal nizations will benefit the most and few of these improved -finance-apps/. opportunities will pass on to individuals generating these 13. “Analog complements” are thoroughly discussed in World Bank (2016, 2). As it notes, “to get the most out of the digital revolution, vast troves of data. If the data economy does not become countries also need to work on the ‘analog complements’—by more inclusive, with wider access to the digital tools and the strengthening regulations that ensure competition among busi- skills to use them, it is likely that the data economy will not nesses, by adapting workers’ skills to the demands of the new benefit the poor. economy, and by ensuring that institutions are accountable.” Finally, as noted, better data will only go so far to improve 14. See http://www.cepal.org/publicaciones/xml/5/48385/leo2013 opportunity; institutions, infrastructure, and rules will need _ing.pdf. 64 Information and Communications for Development 2018 15. For example, see Dave 2017. 25. See https://ec.europa.eu/info/law/law-topic/data-protection 16. See, for example, (a) Universal Declaration of Human Rights, /reform/rights-citizens/my-rights/can-i-ask-company-organi Article 12 (United Nations, 1948), (b) Convention for the sation-stop-processing-my-personal-data_en. Protection of Human Rights and Fundamental Freedoms, 26. See the Data Privacy Project at https://dataprivacyproject Article 8 (European Court of Human Rights, 1950), http:// .org. www.coe.fr/eng/legaltxt/5e.htm, (c) Convention for the 27. See https://support.google.com/websearch/answer/4540094 Protection of Individuals with Regard to Automatic Process- ?co=GENIE.Platform%3DDesktop&hl=en. ing of Personal Data, ETS No. 108 (Council of Europe, 1981), 28. See https://en.wikipedia.org/wiki/Google_Personalized http://www.coe.fr/eng/legaltxt/108e.htm, (d) International _Search. Covenant on Civil and Political Rights (United Nations, 1966), http://www.hrweb.org/legal/cpr.html, and (e) Regula- 29. See https://www.mygov.in/sites/default/files/user_comments tion (EU) 2016/679 of the European Parliament and of the /Digital%20India-Agriculture.pdf. Council of 27 April 2016 on the Protection of Natural Persons 30. See https://datacoup.com/. with Regard to the Processing of Personal Data and on the 31. One survey found an average of more than 100 online Free Movement of Such Data, and Repealing Directive 95/46/ accounts per email address in 2015, and the average number EC (General Data Protection Regulation), to become effective of forgotten passwords to be 37; see Le Bras 2015. May 25, 2018. 32. See the ITU World Telecommunication/ICT Indicators 17. See http://www2.mitre.org/public/jsmo/pdfs/03-01-employer Database. -liability.pdf. 33. See https://protonmail.com/security-details for details. 18. For information, see https://www.theguardian.com 34. See https://www.openoffice.org/privacy.html for details. / technology /2018/mar/17/facebook-cambridge- analytica -kogan-data-algorithm; https://slate.com/technology/2018/04 35. See https://www.mozilla.org/en-US/privacy/firefox/. /the-cambridge-analytica-scandal -suggests-algorithms 36. For discussions on this topic, see http://www.europarl -are-the-new-campaign-donation.html; https://www .europa.eu/RegData/etudes/BRIE/2017/607323/IPOL .newamerica.org/future-tense/events/the-tyranny-of _BRI(2017)607323_EN.pdf; https://www.eff.org/deep -algorithms/. links/2018/04/platform-censorship-wont-fix-internet; and 19. It appears that even when many users are provoked, scale does http://www.oecd.org/competition/rethinking-antitrust-tools not materialize. For instance, following the revelations of the -for-multi-sided-platforms.htm. Cambridge Analytica scandal, a “#DeleteFacebook” move- 37. See https://www.pcmag.com/article2/0,2817,2407617,00.asp. ment got significant coverage in the media, but had limited impact. As of May 2018, user growth continues at Facebook and advertising revenues also increased. See Hsu (2018); References Romano (2018); Murdock (2018); and Sloane (2018). Accenture. 2015. “Guarding and Growing Personal Data Value.” 20. See http://ec.europa.eu/justice/data-protection/article-29 https://www.accenture.com/t20150821T065218__w__ / us /documentation/opinion-recommendation/files/2014 -en/_acnmedia/Accenture/Conversion-Assets/DotCom /wp228_en.pdf. /Documents/Global/PDF/Dualpub_15/Accenture-Guarding 21. Separating personal data from direct identifiers so it is not -and-Growing-Personal-Data-Value-Narrative-Repo. possible to identify the individual (see https://iapp.org/news/a APEC (Asia Pacific Economic Cooperation). 2015. The APEC /top-10-operational-impacts-of-the-gdpr-part-8-pseudo Privacy Framework. Singapore: APEC Secretariat. nymization/). Bergvinson, David. 2017. “Digital Agriculture Empowers Farm- 22. See https://www.coe.int/en/web/conventions/full-list/- ers.” Business Today, January 15. https://www.businesstoday /conventions/treaty/108. .in/magazine/features/digital-agriculture-empowers-farmers 23. For more information, see https://en.wikipedia.org/wiki /story/242966.html. /Comparison_of_instant_messaging_clients#Secure_messengers; Boffey, Daniel. 2017. “Google Fined Record €2.4bn by EU over and Field 2018. Search Engine Results.” The Guardian, June 27. https://www 24. See European Commission, “Can I Ask A Company or .theguardian.com/business/2017/jun/27/google-braces-for Organization to Stop Processing My Personal Data?” at -record-breaking-1bn-fine-from-eu. https://ec.europa.eu/info/law/law-topic/data-protection Brandom, Russell. 2017. “Apple’s New Anti-tracking System /reform/rights-citizens/my-rights/can-i-ask-company Will Make Google and Facebook Even More Powerful.” The -organisation-stop-processing-my-personal-data_en. Verge, June 6. https://www.theverge.com/2017/6/6/15747300 People and Data 65 /apple-safari-ad-tracking-cookie-blocker-google-facebook www.telegraph.co.uk/technology/2018/04/23/duckduckgo -privacy -tiny-private-search-engine-standing-google/. Caldwelladr, Carole, and Emma Graham-Harrison. 2018. Green, Andy. 2018. “Canada’s PIPEDA Breach Notification Regula- “Revealed: 50 Million Facebook Profiles Harvested for tions Are Finalized!” Varonis, May 3. https://blog.varonis.com Cambridge Analytica in Major Data Breach.” The Guardian, /canadas-pipeda-breach-notification-regulations-finalized/. March 17. https://www.theguardian.com/news/2018/mar/17 The Guardian. n.d. “The Cambridge Analytica Files.” https://www /cambridge-analytica-facebook-influence-us-election. .theguardian.com/news/series/cambridge-analytica-files. Chevalier, Judith, and Dina Mayzlin. 2006. “The Effect of Gunawan, Imana. 2017. “App Helps Indonesian Capital Get Word of Mouth on Sales: Online Book Reviews.” Journal ‘Smart’ to Improve Public Services.” Humanosphere, January of Marketing Research 43 (3): 345–54. https://msbfile03 6. http://www.humanosphere.org/social-business/2017/01 .usc.edu/digitalmeasures/mayzlin/intellcont/chevalier /app-helps-indonesian-capital-get-smart-to-improve-public _mayzlin06-1.pdf. -services/. COE (Council of Europe). 1981. “Convention for the Protection Hsu, Tiffany. 2018. “For Many Facebook Users, a ‘Last Straw’ of Individuals with Regard to Automatic Processing That Led Them to Quit.” New York Times, March 21. https:// of Personal Data.” ETS No. 108. http://www.coe.fr/eng www.nytimes.com/2018/03/21/technology/users-abandon /legaltxt/108e.htm. -facebook.html. Confessore, Nicholas, Gabriel J. X. Dance, Richard Harris, and Kalia, Amul. 2017. “Here’s How to Protect Your Privacy from Your Mark Hansen. 2018. “The Follower Factory.” New York Times, Internet Service Provider.” Electronic Frontier Foundation, January 27. https://www.nytimes.com/interactive/2018/01/27 April 3. https://www.eff.org/deeplinks/2017/04/heres-how /technology/social-media-bots.html. -protect-your-privacy-your-internet-service-provider. Data Privacy Project. 2018. “Code of Conduct: Data Science Code Kelly, Michael, and David Satola. 2017. “The Right to Be Forgot- of Professional Conduct.” http://www.datascienceassn.org ten.” University of Illinois Law Review 1. https://papers.ssrn /code-of-conduct.html. .com/sol3/papers.cfm?abstract_id=2965685. Dave, Paresh. 2017. “Credit Giant Equifax Says Social Secur- Le Bras, Tom. 2015. “Online Overload—It’s Worse Than You ity Numbers, Birth Dates of 143 Million Consumers May Thought.” Infographic, Dash Lane Blog, July 21. https://blog Have Been Exposed.” Los Angeles Times, September 7. http:// .dashlane.com/infographic-online-overload-its-worse-than www.latimes.com/business/technology/la-fi-tn-equifax-data -you-thought/. -breach-20170907-story.html. Leber, Jessica. 2013. “In a Data Deluge, Companies Seek to Fill Davenport, Thomas H., Don Cohen, and Al Jacobson. 2005. a New Role.” MIT Technology Review, May 22. https://www “Competing on Analytics.” Babson Executive Education: .technologyreview.com/s/513866/in-a-data-deluge-companies Knowledge Research Center. http://www.babsonknowledge -seek-to-fill-a-new-role/. .org/analytics.pdf. Lehtiniemi, Tuukka. 2017. “Personal Data Spaces: An Intervention Deloitte Review. 2013. “Data as the New Currency: Govern- in Surveillance Capitalism?” Surveillance and Society 15 (5). ment’s Role in Facilitating the Exchange.” Deloitte Review https://ojs.library.queensu.ca/index.php/surveillance-and 13. https://deloitte.wsj.com/riskandcompliance/files/2013/11 -society/article/view/6424. /DataCurrency_report.pdf. Lewis, Paul. 2017. “‘Our Minds Can Be Hijacked’: The Tech Dewey, Catherine. 2016. “98 Personal Data Points That Facebook Insiders Who Fear a Smartphone Dystopia.” The Guardian, Uses to Target Ads to You.” Washington Post, August 18. October 6. https://www.theguardian.com/technology/2017 https://www.washingtonpost.com/news/the-intersect /oct/05/smartphone-addiction-silicon-valley-dystopia. /wp/2016/08/19/98-personal-data-points-that-facebook-uses Lomas, Natasha. 2017. “Form an Orderly Queue! Google Wants -to-target-ads-to-you/?utm_term=.8e8b6b4fb92b. Your Blood (and Other Bodily Fluids). Oh and Your Medical Elgan, Mike. 2016. “Artificial Intelligence Needs Your Data, All of Records.” TechCrunch. https://techcrunch.com/2017/04/21 It.” ComputerWorld, February 22. https://www.computerworld /form-an-orderly-queue-google-wants-your-blood-and .com/article/3035595/emerging-technology/artificial-intellige -other-bodily-fluids-oh-. nce-needs-your-data-all-of-it.html. Lyons, Elizabeth, Zakkoyya Lewis, Brian Mayrsohn, and Jennifer EU (European Union). 2018. “The EU General Data Protection Rowland. 2014. “Behavior Change Techniques Implemented Regulation (GDPR).” https://www.eugdpr.org. in Electronic Lifestyle Activity Monitors: A Systematic Field, Matthew. 2018. “DuckDuckGo: The Private Search Engine Content Analysis.” Journal of Medical Internet Research 16 (8). Standing Up to Google.” Technology Intelligence, April 23. https:// http://doi.org/10.2196/jmir.3469. 66 Information and Communications for Development 2018 Madrigal, Alexis C. 2012. “I’m Being Followed: How Google—and and Barriers.” PLOS, February 2. http://journals.plos.org 104 Other Companies—Are Tracking Me on the Web.” The /plosmedicine/article?id=10.1371/journal.pmed.1001953. Atlantic, February 29. https://www.theatlantic.com/technology Reuters Market Light. 2015. “Addressing Farmers Problems /archive/2012/02/im-being-followed-how-google-151-and-104 through Digital India Initiative.” https://www.mygov.in/sites -other-companies-151-are-tracking-me-on-the-web/253758/. /default/files/user_comments/Digital%20India-Agriculture Magna Global. 2017. “Advertising Forecasts: Winter Update .pdf. (Dec 4, 2017).” https://www.magnaglobal.com/wp-content Richter, Wolf. 2018. “Online Ad Revenues Are Surging, but /uploads/2017/12/MAGNA-Global-Forecast_Winter-Update 2 Companies Are Getting Most of the Spoils.” Business _Final.pdf. Insider, April 27. http://www.businessinsider.com/online-ads Marshall, Aarian. 2017. “The Secret Uber Data That Could Fix -revenues-going-to-google-and-facebook-the-most-2017-4. Your Daily Commute.” Wired, February 3. https://www.wired Romano, Aja. 2018. “How Facebook Made It Impossible to .com/2017/02/ubers-coughing-data-nyc-fix-commute/. Delete Facebook.” VOX, March 22. https://www.vox.com McKinsey Global Institute. 2011. Big Data: The Next Frontier /culture/2018/3/22/17146776/delete-facebook-difficult. for Innovation, Competition, and Productivity. https://www Roosendaal, Arnold, Marc van Lieshout, and Anne Fleur van .mckinsey.com/~/media/McKinsey/Business%20Functions Veenstra. 2014. “Personal Data Markets.” TNO Innovation for /McKinsey%20Digital/Our%20Insights/Big%20data%20 Life. http://publications.tno.nl/publication/34612412/riZsP9 The%20next%20frontier%20for%20innovation/MGI_big /TNO-2014-R11390.pdf. _data_full_report.ashx. Rosenwald, Michael. 2015. “The Digital Media Industry Needs Mineo, Liz. 2017. “On Internet Privacy, Be Very Afraid.” Harvard to React to Ad Blockers . . . or Else.” Colombia Journalism Gazette, August 24. https://news.harvard.edu/gazette Review, September/October. https://www.cjr.org/business /story/2017/08/when-it-comes-to-internet-privacy-be-very _of_news /will_ad_blockers_kill_the_digital_media_indus -afraid-analyst-suggests/. try.php. MIT Technology Review. 2014. “Researchers Test Personal SAS Institute. n.d. “Big Data Analytics.” https://www.sas.com Data Market to Find Out How Much Your Information Is /en_us/insights/analytics/big-data-analytics.html. Worth.” July 9. https://www.technologyreview.com/s/528866 Seller, Marianne, and Paul Gray. 1999. “A Survey of Database /researchers-test-personal-data-market-to-find-out-how Marketing.” Center for Research on Information Technology -much-your-information-is-worth/. and Organizations, University of California at Irvine. https:// Morey, Timothy, Theodore Forbath, and Allison Schoop. 2015. escholarship.org/uc/item/36z642kj. “Customer Data: Designing for Transparency and Trust.” Sengupta, Anirban, and Steven N. Wiggins. 2014. ”Airline Pricing, Harvard Business Review, May. Price Dispersion, and Ticket Characteristics On and Off the Murdock, Jason. 2018. “Mark Zuckerberg Says ‘Delete Face- Internet.” American Economic Journal: Economic Policy 6 (1): book’ Protest Had No Meaningful Impact on His Busi- 272–307. ness.” Newsweek, April 5. http://www.newsweek.com Shu, Catherine. 2018. “Waze Signs Data-Sharing Deal with /zuckerberg-says-deleting-facebook-has-no-meaningful AI-Based Traffic Management Startup Waycare.” Tech -impact-his-business-872876. Crunch, April 26. https://techcrunch.com/2018/04/26/waze Naone, Erica. 2011. “How Useful Is Personalized Search?” MIT -signs-data-sharing-deal-with-ai-based-traffic-manage Technology Review, April 11. https://www.technologyreview ment-startup-waycare/. .com/s/423596/how-useful-is-personalized-search/. Simonite, Tim. 2014. “Sell Your Personal Data for $8 a Month.” Neal, Dave. 2016. “Swedish Deseat Software Will Help You Delete MIT Technology Review, February 12. https://www.technol Yourself from the Internet.” The Inquirer, November 25. ogy review.com/s/524621/sell-your-personal-data-for-8-a https://www.theinquirer.net/inquirer/news/2478419/swedish -month/. -deseat-software-will-help-you-delete-yourself-from-the Singer, Natasha. 2012. “Mapping, and Sharing, the Consumer -internet. Genome.” New York Times, June 17. https://www.nytimes OECD (Organisation for Economic Co-operation and Develop- .com/2012/06/17/technology/acxiom-the-quiet-giant-of ment). 2013. “The OECD Privacy Framework.” OECD, -consumer-database-marketing.html. Paris. http://www.oecd.org/internet/ieconomy/oecd_privacy Sloane, Garett. 2018. “What Facebook Crisis? Ad Sales Skyrocket _framework.pdf. to $11.8 Billion, Users Don’t #delete.” AdAge, April 25. http:// Piwek, Lukasz, David A. Ellis, Sally Andrews, and Adam Johnson. adage.com/article/digital/privacy-concerns-facebook-ad 2016. “The Rise of Consumer Health Wearables: Promises -sales-hits-11-8-billion/3s13271/. People and Data 67 UNCTAD (United Nations Conference on Trade and Development). Whitler, Kimberly A. 2016. “The Personal Data Revolution: Why It’s 2017. Information Economy Report: Digitalization, Trade Time for Marketers to Care.” Forbes, September 18. https://www and Development. Geneva: UNCTAD. http://unctad.org/en .forbes.com/sites/kimberlywhitler/2016/09/18/the-personal-data /PublicationsLibrary/ier2017_en.pdf. -revolution-why-its-time-for-marketers-to-care/#488b46db5681. Upwork. 2018. “Upwork Releases Q4 2017 Skills Index, Ranking World Bank. 2016. World Development Report: Digital Dividends. the 20 Fastest-Growing Skills for Freelancers.” Press release, Washington, DC: World Bank. February 7. https://www.upwork.com/press/2018/02/07 World Economic Forum. 2011. “Personal Data: The Emergence of /q4-2017-skills-index/. a New Asset Class.” World Economic Forum, Geneva. http:// Wang, Teresa, and Malay Gandhi. 2014. “The Future of www3.weforum.org/docs/WEF_ITTC_PersonalDataNew Biosensing Wearables: A Review of the Current Landscape Asset_Report_2011.pdf. and Future of Biosensing Wearables.” Rock Health. https:// ———. 2013. “Unlocking the Value of Personal Data: From rockhealth.com /reports/the-future-of-biosensing-wear Collection to Usage.” World Economic Forum, Geneva. http:// ables/. www3.weforum.org/docs/WEF_IT_UnlockingValuePersonal Warren, Mark. 2016. “The Cure for Cancer Is Data—Mountains Data_CollectionUsage_Report_2013.pdf. of Data.” Wired, October 19. https://www.wired.com/2016/10 Zannier, Federico. n.d. “A Bite of Me.” https://www.kickstarter /eric-schadt-biodata-genomics-medical-research/. .com/projects/1461902402/a-bit-e-of-me/description. 68 Information and Communications for Development 2018 Chapter 5 Firms and Data Introduction Digital platforms D ata is becoming crucial to the competitiveness Defining digital platforms of nations, regions, cities, and companies. This Digital platforms might be defined as “multisided chapter highlights the implications of the tran- marketplaces with business models that enable producers sition to a data-driven economy for firms, especially in and users to create value together by interacting with each emerging markets. The emergence of the data economy other” (Still et al. 2017), and by facilitating matching, has stimulated development of new business models and searching, exchanging, transactions, and so on (Evans 2013a, is transforming many of the key functions of private sector 2013b). Marketplaces rely on information to adjust prices firms, including product development, customer relation- and impose rational order, but information is frequently ships, supply chains, and core enterprise functions such as uneven and incomplete. Digital platforms offer advantages marketing, human resources, and finance. Governments in over traditional marketplaces through scale and network emerging markets are keen to explore and implement the effects that increase the information flow and interaction right policies to allow private sectors to enhance competi- between participants. Participants derive benefits from tiveness and benefit from new opportunities in the digital communications networks that increase as others join economy while mitigating risks. the system (Rohlfs 1974). For marketplaces that bring The chapter examines the impact of data on firms from together suppliers (or advertisers) and users (or informa- three distinct perspectives. The first part of the chapter intro- tion consumers), economies of scale become even more duces an assessment framework for digital platforms—a important (Rochet and Tirole 2003). Multisided platforms dominant aspect of data economies—and features case benefit from positive network externalities, as the utility studies of selected digital platforms in emerging markets. of each side increases as participants increase on the other Next, the chapter looks at how data affects firms, highlight- side. For example, the utility of a car-sharing platform for ing the tension between data as an equalizer, but also as a key each side increases with the increase of drivers and riders. competitive differentiator. Finally, the chapter looks at how The scale effects are not uniformly positive, however, and data affects small and medium enterprises (SMEs)—and policy makers must recognize risks such as dominance and their specific needs. anticompetitive behavior. 69 Digital platform enablers Platform enablers have important implications for Digital platforms typically combine physical (and virtual) economic development. Emerging and transitioning econ- and behavioral (and market) enablers. Physical enablers omies often lack pervasive broadband internet infrastructure (figure 5.1) include digital infrastructure (fixed and mobile (Kelly and Rossotto 2012), and present wide disparities in broadband networks), smartphones, payment tools, internet access among population groups (these differences geolocation, cloud-based services, security, and ancillary relate to, among other things, urban versus rural, gender, enablers (such as distribution, logistics, and intermediary age, education, and income differences). Equally important goods). Behavioral or market enablers (figure 5.2) nudge divides affect access to devices such as smartphones and consumers toward buying goods or accessing services in laptops (World Bank 2016). Overcoming the digital divide a peer-to-peer economy in which platforms increasingly is thus essential to developing digital platforms in emerging mediate interactions, typically coordinated by peer-based markets. In the Middle East and North Africa region, the trust relationships. This development is sometimes called ride-hailing platform Careem emphasizes the social value collaborative consumption (Constantinou, Marton, and proposition of not only creating jobs, but fostering social Tuunainen 2017). value by allowing drivers to become micro-entrepreneurs, including by equipping them with smartphones. The development of the physical and virtual enablers of Figure 5.1 Physical and virtual enablers digital platforms in developing countries may require dedi- cated policies, technical assistance, and investment. Digital infrastructure may also require a combination of telecom- Cloud- based munication market liberalization, regulatory reform, and services and better targeting of subsidies to extend the commercial viabil- security Digital DIGITAL payments ity of broadband infrastructure (Kelly and Rossotto 2012) or PLATFORMS Broadband public-private partnerships (Ragoobar, Whalley, and Harle Multisided 2011). Increasingly, as shown in chapter 3, these physical PHYSICAL AND platform Geolocation enablers are software based, such as artificial intelligence, VIRTUAL ENABLERS services Consumer the Internet of Things (IoT), machine learning, and autono- equipment mous vehicles (Lal Das et al. 2017; Schwab 2017). They may Distribution and also require harmonized data protection and privacy stan- logistics dards to facilitate the development of cross-border oper- ations (ITU 2015). The development of technology enablers has been crucially important, for instance, for Alibaba’s ecosystem development (Tan et al. 2015) to bring together Figure 5.2 Market and behavioral enablers the trading platform, payment system, and logistics network that forms the basis for its e-commerce business platform (Tsai 2016; see figure 5.3). Peer-to- peer feedback- The study of the market and behavioral enablers of mediated relationship digital platforms constitutes a research agenda by itself. In Collaborative many cases, participant behavior and market development DIGITAL consumption “Long-tail” PLATFORMS marketing in emerging markets closely mirror that in high-income Multisided markets. For example, digital platforms to match labor MARKET AND Buying platform supply and demand are popular in emerging markets: the BEHAVIORAL access and ENABLERS “servitized” Single- Philippines, the Russian Federation, and Ukraine are among products versus Network multi-homing the top 10 countries providing skilled labor on Upworks’ externalities digital platform. Alibaba is a serious global competitor for and critical mass eBay and Amazon, and Alipay’s transactions are a multiple of those of Paypal. 70 Information and Communications for Development 2018 Figure 5.3 Geographical concentration of digital multinational enterprises with revenue in excess of US$1 billion, by region, 2016 North America Asia Europe Africa and Latin America Intel Facebook SAP Baidu Naspers Yahoo! Flipkart Google Alibaba Spotify JD.com Rakuten Airbnb Amazon Naver Microsoft Tencent Softbank Uber Xiaomi Netflix Apple Oracle Snapchat Paypal Salesforce Priceline 63: $2.8 trillion 42: $670 billion 27: $161 billion 3: $61 billion Publicly listed Privately owned Source: UNCTAD 2017, citing Van Alstyne 2016. But ability to scale up and reach critical mass is limited specific product preferences, or a large number of products to a relatively few emerging markets. Plus, platforms that sell in small quantities (Enders et al. 2008). So-called often exclude many economic actors, such as consum- long-tail effects on digital platforms give users more choice, ers outside the reach of mobile broadband coverage enabling them to search for less common items or services or without smartphones and SMEs without access to from foreign countries, such as Indian music (Booth 2017) technology. SME owners can be encouraged to partici- or Latin American cultural artifacts (Suominen 2017). The pate in platforms through tax breaks or subsidies or be value to advertisers of capturing long-tail marketing data given training or access to technology (Badran 2014). fosters marketplaces, such as Jumia in Africa (see box 5.1) Incentives to global platforms to localize businesses by and MercadoLibre in Latin America. Other differentiators partnering with local businesses could also be an option, may include price discrimination, delivery, geographic reach, as shown, for example, by the Uber-Yandex agreement in and wholesale versus retail (Täuscher and Laudien 2018). Russia. Platforms in emerging markets share many of the char- acteristics of business models encountered in high-income markets. Regional matching platforms like Laimoon.com Business models for digital platforms (matching labor supply and demand in Arab countries) Emerging business models for multisided platforms are and Arabmatrimony.com seem to mirror the models, based around the bargaining power of different partici- respectively, of Upworks and Match.com, taking cultural pants to drive revenues and organize dispersed information differences into account. Successful platform models that to make it available to market participants (Rochet and worked in advanced economies were also adopted by Tirole 2003). By using platforms, firms can drastically slash local platforms, as the case of Careem shows. Careem is transaction costs, creating new markets (Henten and Wind- a transport network company based in Dubai, with oper- ekilde 2016). By using social networks combined with such ations in 53 cities in the Middle East, North Africa, digital platforms, firms can leverage a “long tail” of market and South Asia. The company was valued at about participants, that is a large number of customers with US$1 billion as of 2017. Firms and Data 71 Box 5.1 Jumia: “Cash on delivery” e-commerce in Africa Jumia is an African online shopping website, primarily for electronics and fashion goods, developed through a strategy of envelopment and service diversification. Jumia launched in Lagos, Nigeria, in June 2012 with initial funding from Rocket Internet (Germany), MTN (South Africa), and Millicom (Luxembourg). It has grown through acquisitions and foreign expan- sion and, in fiscal year 2017, it generated some €93.8 million in revenue (US$111.4 million), although it continues to post substantial losses. It had 2.2 million customers across 14 countries in Africa, though it also offers sales elsewhere.a As in most of Africa, Nigeria mostly uses cash, with 99 percent of transactions cash related. When launched, according to MasterCard,b some 59 percent of the Nigerians questioned the safety of online transactions and 43 percent were concerned about the quality of the products delivered and would rather buy from stores where they could physically inspect products. To address these constraints, Jumia uses a cash-on-delivery mode of payment. Consum- ers can pay by cash or with a point-of-sale terminal and receive their receipt on the spot. This method addressed consumers’ concerns, giving them the human contact associated with visiting a physical shop while building customer trust as well. Some earlier e-commerce ventures, such as kasuwa.com and sabunta.com, had failed and the incubator investor, Rocket Internet, merged and rebranded them as Jumia in 2012. Just a year after introducing the cash-on-delivery option in six Nigerian cities, Jumia.com.ng had become the most popular and fastest-growing online merchant in the country and accounted for 92 percent of total orders.c In the Kenyan market, which Jumia entered in March 2013, the company also offers the facility to pay through mobile money. In addition, it has launched offline “experiential centers” to help consumers overcome their doubts about the look and feel of the products they are purchasing.d A major constraint on home delivery is the lack of formal street addresses in many African cities, or detailed mapping, and Jumia has had to develop its own maps for delivery. For the moment, the global e-commerce majors, such as Amazon and Alibaba, have not developed a strong presence in Africa, though Amazon is building a fulfillment center in Cape Towne and Alibaba has established a fund for young African entrepreneurs.f It remains to be seen whether local African e-commerce companies, such as Takealot (South Africa), Kilimall (Kenya), Konga (Nigeria), as well as Jumia, can build sufficient scale to resist the global majors when they do arrive. a. Financial information extracted from Rocket Internet financial report at https://www.rocket-internet.com/sites/default /files/investors/2017_FY%20Rocket%20Internet%20and%20Selected%20Companies%20Results.pdf. b. Mastercard 2012. c. Jumia, “Africa’s Jumia and Zando Receive Significant Funding from J. P. Morgan Asset Management,” Press Release, October 9, 2012 (https://blog.jumia.com.ng/wp-content/uploads/2012/10/J.P.-Morgan-Invests-in-Jumia2.pdf). d. TechMoran 2017. e. BusinessTech 2018. f. See http://disrupt-africa.com/2017/07/alibabas-jack-ma-launches-10m-african-young-entrepreneurs-fund/. As platforms have evolved, four main business models platforms provide to their sellers based on their trans- have emerged for revenue: action history on the platform) • Commission-based revenue Täuscher and Laudien (2018) introduce a taxonomy • Subscription-based and service-based of platform business models in which, in addition to the four revenue models, other dimensions of the revenue • Advertisement-based source are also captured (such as price discrimination and • Tertiary services developed based on the data from the source), in addition to the delivery dimensions (consumer network (for example, supplier financing that many to consumer, business to business, business to consumer, 72 Information and Communications for Development 2018 global, regional, and local, among others). Digital platforms “critical mass” factors, and reversibility of participation may be primarily wholesale (business to business) or retail create entry barriers and are likely to be more pronounced in (business to consumer). Platforms may mutate the scope developing countries (Evans and Schmalansee 2016). and breadth (reach, timeliness) of markets or lower entry Competition among multisided platforms depends on barriers (or both), thereby affecting competitive dynamics. several factors, including network effects, and “single-homing” Organizational science has introduced additional elements, versus “multi-homing” scenarios (Armstrong 2006). If an agent adding new dimensions to this taxonomy, including the degree uses only one platform, the agent is said to be single-homing; of standardization of output (product) and (input) and how if the agent uses multiple multisided platforms, the agent is they shape different organizational mechanisms. multi-homing. Network effects may lead to situations in which The biggest obstacle to the development of certain digital a proprietary platform may be socially desirable, as it partially platform models is the relative underdevelopment of the internalizes two-sided indirect network effects and direct advertising industry in many emerging markets. This factor competitive effects on the producer side. Ruutu, Casey, and may limit the ability of platforms in those markets to subsidize Kotivirta (2017, 128) indicate that “if platform adopters are a side of the business and limit the range of possible business able to react quickly, achieving a critical mass may be difficult models. One outcome is to encourage the development of because the platform firms cannot accumulate enough resour- “transactional” models to the detriment of commission-based ces for sufficient platform development.” platforms. A real estate owner in a high-income country, for Competitive dynamics can be altered by “platform envel- instance, can choose whether to list a property on a pure, opment,” a strategy through which an entrant platform can local marketplace (whose business model is purely advertising rapidly gain market share by entering another platform based) or to use a foreign, commission-based platform like market and harnessing its network effects by offering a Airbnb. But absent a mature advertising market, the consumer multiplatform bundle (Eisenmann, Parker, and Van Alstyne may have a more limited choice of platforms. 2011). An example would be Google’s offer of various services around its core search platform, including Google Translate (translation software), Google Checkout (online Digital platform dynamics payment), Chrome (browsing), and Google Docs (produc- Platform dynamics have business and policy implications. tivity software). The services are often offered free of charge The dominance of a platform, arising from network effects to the user (that is, paid by advertising), whereas other (winner-takes-all), raises competition concerns, although competitive service offerings may require payment. some argue that dominance can benefit consumers through The initial growth phase can be accelerated through open greater convenience. Often, the achievement of critical interfaces (that reinforce cross-side network effects), as well mass drives platform dynamics (Evans 2013a, 2013b; Ruutu, as by the ability to transfer user data among competing plat- Casey, and Kotivirta 2017). forms. These tactics may lead to the envelopment of local Some authors propose a system dynamics simulation service offerings in developing countries. They thus raise model of platform competition (Ruutu, Casey, and Kotivirta competition policy concerns as local companies in these 2017), highlighting three cases. In the “chicken-and-egg” countries often lack economies of scale to respond in kind. scenario, no platform achieves critical mass. In the “winner- This may explain, in part, why the largest internet companies takes-all” scenario, a vendor locks the participants into one are so clustered in the United States and Asia (see figure 5.3). dominant platform. The final scenario—”winner takes Different dynamics can emerge and mutate over time. In some”—is characterized by the “collaboration and competi- the U.S. market for online platforms for books, for example, tion scenario in which several platforms coexist in balanced eBay acquired half.com, eliminating its most direct competitor. competition” (Ruutu, Casey, and Kotivirta 2017, 128). However, it did not prevent other platforms, notably Amazon Various models have focused on the conditions for .com (new and second-hand books, and a wide range of other multiple platforms to grow first, and then to coexist in a products and services), from dominating the same market competitive market. Both considerations matter in develop- segment, deploying a different business model. ing countries, and have policy and regulatory implications Will competition among platforms in emerging markets (Kenney and Zysman 2016; Frieden 2017). Network effects, follow a similar dynamic? The choices may be more limited. Firms and Data 73 Network effects and critical mass considerations may skew three main drivers of platform dynamics: network effects, competition toward foreign platforms. The huge size of the localization, and envelopment. To some extent, it will also domestic Chinese e-commerce market was a factor leading determine whether winner-takes-some or winner-takes-all to the emergence of Alibaba and Alipay as global leaders trends dominate. (Tsai 2016), while the ability to cater to domestic demand allowed Yandex to retain two-thirds of the addressable Firms in the data economy market in Russia. But few developing countries have such economies of scale, and there are many cases in which local The Organisation for Economic Co-operation and Develop- platforms do not emerge as winners. As noted, the fear of a ment (OECD), in its submission to a Group of Twenty winner-takes-all scenario dominated by foreign platforms conference, provides useful background about the oppor- seems to be a major concern in many markets. tunity of the data economy: “As the cost of data collection, Partnership is another business strategy undertaken by storage and processing continues to decline dramatically, digital platforms, especially in emerging markets. Incentives ever larger volumes of data will be generated from the for global platforms to localize their businesses by partnering IoT, smart devices, and autonomous machine-to-machine with a local business partner can also be an option. In China, communications” (OECD 2017, 63). This will require a new Uber decided a better strategy was to sell out to its local approach to thinking about infrastructure in the twenty-first partner, Didi Chuxing (Kharpal 2016), and in Southeast Asia century, with the definition expanded to encompass broad- it is selling to Grab in exchange for a stake in the combined band networks, cloud computing, and data itself, which company (Sherman 2018). The common shareholding in all drives productivity growth (OECD 2017). three companies of Japan’s Softbank, now the Visions Fund, In the United States, for instance, Brynjolfsson, Hitt, seems to have played a part here (The Economist 2018a). and Kim (2011) estimate that output and productivity in Access to user data is of vital value for online platforms to firms that adopt data-driven decision-making are 5 percent keep advertisers onboard and is a crucial tool for competing. to 6 percent higher than what would be expected from Owning the end customer’s data is akin to owning the their other investments in and use of information and market. This is one reason traditional freight forwarders communication technologies (ICTs). A study of 500 firms are being squeezed by vertically integrated e-commerce in the United Kingdom found that firms in the top quar- companies.1 Online platforms have an interest in locking tile of online data use are 13 percent more productive out rivals that may threaten their market dominance. This than those in the bottom quartile (Bakhshi, Bravo-Biosca, makes data portability—the ability to transfer a user’s data and Mateos-Garcia 2014). Overall, these firm-level studies from one platform to another—a critical policy issue (Graef, suggest that firms’ use of data and data analytics raises labor Wahyuningtyas, and Valcke 2015). Data portability benefits productivity faster, by 5 percent to 10 percent (OECD 2015). users and secondary players in the market, but will most Other studies (such as Täuscher and Laudien 2018) likely be opposed by the dominant players. identify several characteristics of digital business that create “dynamic competition and high consumer surplus” Digital platforms in developing countries (Täuscher, n.d., 10). Many of these characteristics depend on A study of digital platforms in emerging markets provides data as their fundamental lever. significant insights into business leaders and policy makers. Broadband and smartphone access will have a direct Product and service design impact on network effects and platform diffusion; the In many industries, data has become the new product, rather maturity of the advertising market can exclude or boost than the physical goods that firms traditionally sold. When advertising-based platform models; conversely, the maturity you buy a custom-fitted suit, you often become an unwitting of digital payments, such as mobile money, in emerging participant in a data economy in which “clothing companies markets may allow firms to determine the development now see body measurements (data) as one of their most of transaction-based models. The interplay of rapidly prized currencies” (Harwell 2018). Stitch Fix, for example, changing enabling conditions in emerging markets and which had nearly US$1 billion in sales last year, is really a rapidly changing business models (figure 5.4) will affect the data company in disguise (it gathers dozens of data points 74 Information and Communications for Development 2018 Figure 5.4 A methodological approach to assessing digital platforms in emerging markets Broadband Value creation Long-tail Network marketing Localization Value effects and delivery envelopment Network externa- Value Peer-to- lities capture Winner takes peer some or winner feedback takes all Physical and business enablers Business model Platform dynamics on each customer, including weight, jobs, and past pregnan- scale of investments required to run data-driven businesses cies). Similarly, the moment you buy a car, you start making has grown substantially, the marketplace has begun to tilt money for companies like Otonomo, which sells driving in favor of large-scale incumbents (Surowiecki 2016). Other data to third parties. The company has raised US$40 million reports have likewise concluded that the rise of big digital in investments (Etherington 2017) designed to “move from businesses may be squelching competition (Casselman 2017) the age of data mobilization, to the age of data monetiza- by using the power of their (data-driven) platforms. tion.” And finally, when the world’s top-ranked tennis player This uneven growth between startups and incumbents Simona Halep fell out with her clothing sponsor before the is not limited to developed countries and there are still very Australian Open “she took to the internet to find a design she few examples of scaled up data-driven firms in the develop- liked, then ordered it from a seamstress in China. Twenty- ing world (see figure 5.3). Firms in the developing world face four hours later, it was in her hands” (Matthey 2018). several additional data-specific challenges that create a tilted Take a simple example from daily life: smartphone speech field in the marketplace: recognition can help write text messages three times faster • Low “datafication” of the economy (for instance, govern- than human typing (Carey 2016), a dramatic improve- ment records and archives may not be digitized) ment over just a few years ago when speech recognition was considered an irritant (or an amusing novelty at best). • Limited data talent pool The availability of more and better data (that feeds artifi- • Restrictive data policies (localization, poorly developed cial intelligence) is the single most important reason for privacy and consumer protection laws) this enhancement, and firms that can successfully utilize the ever-increasing amounts of data at their disposal are • Underdeveloped data ecosystem beginning to separate themselves from their competitors by • Generally, a higher unit price for data relative to afford- delivering new products and services that both depend on ability (see map 5.1) and generate vast amounts of data. If data is to be the new oil, then firms must invest heavily in data refineries and new capabilities (see chapter 2). In Data-driven supply chains 2016, Amazon, Alphabet, and Microsoft together spent nearly Supply chains are a vital way for companies to create value US$32 billion in capital expenditure and capital leases, up by and deliver products and services.2 Technology-based 22 percent from the previous year. Firms are also investing supply chain innovation initially gave firms such as significantly in developing analytical tools that can make Walmart, which invested heavily in radio frequency identi- sense of this data in real time and convert this data into arti- fication chip technology, tremendous competitive advan- ficial intelligence or “cognitive insights.” Unfortunately, as the tage (FlashGlobal 2018). But firms like Amazon, which Firms and Data 75 Map 5.1 Average price of 1 gigabyte of mobile data per month, by country, 2016 a. Percent of per capita GDP 0–2 2.1–5 >5 No data IBRD 43804 | SEPTEMBER 2018 b. US dollars $1.05 $133.70 No data IBRD 43931 | SEPTEMBER 2018 Source: ICTdata.org. Note: The maps use International Telecommunication Union methodology (see table DN.1). have mastered data and digital innovation (Burson 2016), entire supply chain and is being remade significantly as are now showing the way. Koçoƥlu et al. (2011) describe businesses digitalize ever more. integration with customers, integration with suppliers, That said, McKinsey Global Institute (2018) found that and interorganizational integration as the key value driv- as companies have begun to digitize products and services ers in supply chain integration. Information or data rapidly, supply chain digitization has lagged,3 (yet the same sharing (with customers, suppliers, internal functions, firms expect the digitization of supply chains to have the and across organizations) is a core component across the highest impact on revenue in the near future). Progress has 76 Information and Communications for Development 2018 been especially slow in the management of supply chain be argued that several disruptors succeeded because the data, according to McKinsey Global Institute (2017).4 Chal- incumbents were not yet digitally or data savvy). Some of lenges include the development of data infrastructure to these examples include the following: manage vast amounts of data (what Ernst and Young [2016] • Personalization. Incumbent firms can often deliver more called the “out-of-control data growth trap”), the ability personalized products and services to customers given to link disparate sources of data, and the development and the vast amount of data they have collected about them. utilization of tools to analyze data productively. These chal- lenges have been heightened by the growth of data-fueled • Predictive analytics. Firms can use their vast data troves to disruptive technologies—such as the IoT, artificial intelli- predict the movies you like, the books you are likely to buy, gence, robotics, and blockchain—that are fast becoming and your likelihood of trying rival products and services. essential elements of supply chain management technology This gives them a significant advantage against competition. but are frequently beyond the capabilities of SMEs and their • Prescriptive analytics. Data-smart firms are able to react customers. Digital and data technologies that have integrated to events in real time to resolve customer management millions of firms and their suppliers in common global value issues (for instance, vouchers to compensate for a delayed chains are also gradually beginning to separate them. flight [Brahm, Cheris, and Sherer 2016], rather than a Amazon is an illustrative example of a firm that has routine customer survey, for instance). used its mastery of supply chain data to distance itself from Data-poor firms are at an inherent disadvantage in these competitors but also begin to erode the space of its suppliers scenarios. and sellers on the platform. With its granular visibility into the operations of both the buyers and sellers on the platform, SMEs in the data economy Amazon has realized that it can manufacture and distribute many products on its platform cheaper than other suppliers Although the digital economy is increasingly dominated by can (via the Amazon Basics program). Streamlining of the a handful of tech majors, a multitude of innovative SMEs manufacturing, distribution, and retail of these products, nevertheless are the drivers of the mobile and digital industries, combined with its mammoth scale and superior data smarts, particularly in newly emerging market segments such as data gives Amazon tremendous competitive advantage. Can its for self-driving vehicles or mobile applications. Opportunity competitors without access to any equivalent market data, exists therefore for tech-based SMEs to play a major role in such as Jumia (see box 5.1), compete? the data-driven economy. The mobile ecosystem in Nigeria, for instance, was worth an estimated US$8.3 billion in 2017 Marketing and customer relationship management (Boateng et al. 2017), and the digital industry may contribute Customer acquisition, management, and retention are core 7 percent of Mali’s GDP (da Silva 2014). functions of business and digital and data technologies are However, equally important is the impact of the data transforming this landscape. In many ways, data and digital economy on SMEs in nontech sectors. It is estimated are the ultimate equalizing force. Firms using digital tools that half of all job opportunities in middle- and low- and platforms theoretically have equal access to customers income countries are generated by local SMEs (Matthee and around the world (local laws permitting), can use communi- Heymans 2013). As a result, SMEs are instrumental in eradi- cation tools and platforms to stay engaged online in real time, cating poverty, creating economic growth, and empowering and take advantage of a variety of payment systems and plat- citizens to become productive economic agents. But SMEs forms plus logistics services to deliver products and services typically lag larger firms in the adoption of certain digital worldwide. This is how Uber was able to reach riders around technologies (Andrews, Criscuolo, and Gal 2016). Although the world, for instance, and how Instagram became a global SMEs are just as likely to use broadband internet as larger rage. If these firms could acquire millions (and even billions) firms, gaps still exist in adopting more sophisticated digital of customers and scale globally quickly, then so can other tools. The growing complexity of new technologies requires firms if they can create appropriate products and services. investment in new skills, and where these skills are lacking it There is an element of truth to this theory, but data is slowing technology diffusion among smaller and younger confers several advantages on incumbents (indeed, it may firms (Andrews, Criscuolo, and Gal 2016). Firms and Data 77 SME advantages, drivers, and constraints of charges, for instance, is spurring new business models data adoption for SMEs to provide credit to the underserved. Even SMEs try to absorb new technologies and innovation, but are in sectors unrelated to financial products and services, often constrained by limited availability of skilled workers, firms are developing new data-centric business lines and particularly in emerging markets, in turn limiting potential alternative revenue streams out of the data they collect for growth and job creation. Starting in the 1990s, many from customers. Firms in Sub-Saharan Africa, such as SMEs in developing countries began to adopt modern ICTs, M-KOPA Solar (Kenya),6 Off Grid Electric (Tanzania),7 increasing profitability and productivity (Badran 2014). In PEG Africa (Ghana),8 and BBOXX (Côte d’Ivoire),9 addition, ICT made training and education more accessible have not only revolutionized energy access, but are for workers. This could eventually raise the employability of also starting to support financial inclusion. Through low-skilled workers. “pay-as-you-glow” business models, these providers SMEs are characterized by potentially advantageous allow low-income, mostly rural consumers to have solar features that distinguish them from other businesses. Their energy at home. On the basis of the data collected on the relatively small market size allows them to adapt quicker timeliness of repayments they accumulate for the home to changing market conditions, and they are less likely to solar systems they offer, these energy companies can have stranded assets, both of which increase their chances allow customers to build a credit history and thus access of success. Increasing digitization dramatically reduces credit and loans. transaction costs for collecting information, communica- However, challenges to the use of data and data analytics tion, and data controlling (World Bank 2016). Through exist, particularly in emerging markets, which are more easier access to information and the use of complex data acute for SMEs than for larger firms, as discussed below. analytics, firms may analyze the interdependency and First is financial and access constraints. SMEs tend to buying patterns of users to pursue targeted advertisement, have limited access to financial resources, which makes and adjust their inventory accordingly. SMEs can exploit it hard to invest in new technologies and maintain them. low entry barriers to benefit from the potential disruption Limited financial resources also cause SMEs to lack a formal of data on existing models. Moreover, SMEs’ digital tech- risk management practice, even for those that do have an nology adoption barriers can be lowered by the transition information technology department (Priyadarshinee et al. from hard infrastructure investments to platform-based 2017). In addition, SMEs in emerging markets often face digital services. The increasing availability and range of obstacles accessing data relevant to their business. Larger cloud-based tools for enterprise management is particularly firms gain access to that same data, often owned by the relevant to SMEs.5 government, or are able to pay for it from private sources, Data analytics allows firms to establish new forms of thanks to larger financial resources or networks of contacts customer engagement, exploit digital distribution channels, not available to SMEs. and serve new customers. Data analytics, combined with Second is limited awareness. SMEs also tend to lack aware- voice and vision recognition, enables firms to complement ness of the opportunities offered by digitized business and or substitute for human labor with machines (such as auto- operations, which affects their ability to adapt and compete mated call answering and recording to reduce call center in a fast-evolving business environment. A 2014 survey employees). Leveraging data can also affect competition, among 1,000 SMEs in Germany revealed that for 70 percent with SMEs transforming processes, facilitating innova- of enterprises with annual revenue below €500 million, tion, and addressing key challenges. Access to data can the digitization of processes was still seen as irrelevant. revolutionize decision-making with enhanced visibility of Making the situation worse is that many available ICT prod- firm operations and improved performance measurement ucts and information do not necessarily take the specific techniques. needs of SMEs into account. Third, human capital limitations are a constraint. Innovative data-driven business models for SMEs Investments in new technologies often require invest- The use of alternative data to build credit histories by ments in complementary knowledge-based assets. SMEs scanning users’ mobile phones for their history or credit frequently lack the skilled people to benefit from new digital 78 Information and Communications for Development 2018 Box 5.2 Agribusiness SMEs and data-driven supply chains Digital technologies can change farm practices and agricultural structures and, hence, contribute to the prosperity and resilience of farming systems. Agribusiness supply chains are increasingly becoming data driven, which raises the need to move toward higher levels of data integration along production chains. Farmers and agribusinesses can benefit from enhanced data usage for improved sustainability, food safety, resource efficiency, and reduced waste. Over the last decade, information and communication technology (ICT) use in the farm sector has increased significantly. The World Bank (2017) highlights a range of areas in which ICT has been success- fully applied (such as the use of GPS for farm field management, sensor data on crops and cattle to predict diseases, weather data, logistics tracing and tracking, online shops, agricultural market pricing data, and many others). Food supply chain players have been making advanced use of ICTs, with the next steps related to small and medium enterprise (SME) capabilities to unlock the potential generated by ICT applications. Farm data is still hardly shared with sectoral stakeholders, analyzed by intelligent software, or combined in regional analysis and advice. Hence, food supply chains may not fully take advantage of the large amounts of potential data, especially to smallholder farmers (figure B5.2.1). Agribusiness is a sector with many small firms whose need will increase to invest in software and combine it with data seamlessly available to business partners and govern- ment agencies—as large firms already do internally in their enterprise resource planning systems. However, the limited interoperability of data and information systems makes it more complicated. This holds for SME-to-SME and SME-to-government communication as well as SME-to-big-company communication. For instance, consider the challenge for a large avocado cooperative that wishes to exchange digital data with thousands of farmers spread across Peru, or a dairy manufacturer that wants to monitor operational data from Ethiopian farmers. As such, business-to-business digital platform applications, and data common standards, become crucial to foster data usage in heavy supply chain sectors, like agribusiness. Figure B5.2.1 How more data contributes to current business models in the food chain Logistics Input Software Food Retail and Farmer solution industries providers processor consumer providers Transport Transport Transport GRIN Small Cost price Service Cope with retail Loyalty Feed the growing world Sustainability Food safety Health Precision farming: Segment products Better control and input suppliers Better management Benchmarked with decisions competitors Consumer Sophisticated decision support technology, (pre- and postsales) more advice Better service concepts (e.g., in-store replenishment) Source: Poppe et al. 2013. Note: GRIN = Genetics, robotics, information, and nanotechnologies. Firms and Data 79 Box 5.3 Alibaba’s success: SMEs as the foundations of the business model Alibaba, the world’s largest e-commerce platform by sales volume, supports an estimated 10 million jobs, or 1.3 percent of China’s workforce. One of the most valuable assets Alibaba and other e-commerce operators accumulate is data. Data connects small and medium enterprises (SMEs), many of which are in the 2,000 plus so-called Taobao villages,a to Alibaba’s ecosystem, and ultimately to consumers. Each transaction contributes to improved knowledge about the economy and consumer behavior. This information, coupled with data analytics, supports new business lines and product innovation, such as extending credit to small firms based on automated evaluations of creditworthiness (figure B5.3.1). Chinese companies selling on Alibaba, in large part SMEs, reach an average of 3 and in some cases up to 100 different export destinations, up from an average of 1 and a maximum of 50 export destinations for offline firms. Alibaba further guarantees the on-time delivery of money from foreign buyers and has implemented a system to verify sellers on its website for business-to-business transactions. Firms can acquire a “gold” supplier status by paying for a third-party verification company to conduct on-site quality control. Alibaba is promoting its model abroad, with recent memoranda of understanding with both the Malaysian and Mexican govern- ments, to provide SMEs in developing countries the skills to benefit from cross-border trade. Figure B5.3.1 Alibaba’s physical and virtual enablers Computing services Data management services Communicating tools Weibo Marketing Social network affiliates Mobile internet Digital entertainment Mobile browser Market place B2C trading platforms platform C2C platform O2O Alipay Search tool International platform payment Wallet Group-buy services China location-based Logistics data SME loan and Exchanging apps Logistics platform TRADING Payment and finance support PLATFORMS finance platform Ecosystem Ecosystem contributor platform layer layer LOGISTICS PAYMENT Ecosystem infrastructure layer Transaction integrating direction User demand layer Supporting integrating direction BUYERS SELLERS Integration within value hub Sources: Adapted from World Bank 2016 and Tsai 2016. Note: B2C = business-to-consumer; C2C = consumer-to-consumer. a. See https://sampi.co/taobao-villages-china-rural-ecommerce/. 80 Information and Communications for Development 2018 Box 5.4 The app economy in the Arab world According to a recent report by the Mohammed Bin Rashid School of Government, more than 96 percent of users in the Arab States region said they personally had experienced a positive impact from digital platform apps, with some 55 percent saying it saved them time, 33 percent that it saved money, and 8 percent that it had personally generated income from delivering services on sharing economy apps. On the other hand, 3 percent of users reported negative impacts on the income of the users, mainly because these services hurt their existing sources of income (for example, taxi drivers and hotel owners). Digital platforms include transport appli- cations, the most popular type of sharing economy services in the Arab world. Slightly more than half reported using the Careem and Uber apps, and a quarter use accommodation apps, such as Airbnb. Local alternatives, such as Tirhal and Mishwar, were also popular in some countries. Sources: GSMA 2017; Salem 2017. technologies, the resources to train these workers, or the affects the ability of SMEs to adopt digital-data-generating management that can help them make the most of the new tools. These frameworks should include privacy policies, technologies. The lack of availability of skilled labor inhibits intellectual property, data security, and access rights. Emer- the adoption of data analytics, complex data integration, gent practices also risk reducing confidence in the digital and model building in SMEs, especially in developing coun- economy and the incentives to adopt ICT. Discrimination tries. That SMEs in emerging markets have a harder time enabled by data analytics, based on profiling customers by competing for scarce skilled labor against larger firms, both where they live, for example, may create greater efficiencies local and foreign, compounds this challenge. and innovation but also limit individual freedom (OECD Fourth, new data sources may require remodeling of 2015). On the other hand, disruptive technologies that tackle existing systems, such as SME warehouse systems. This is data governance aspects, such as distributed ledger technol- particularly true considering the volume and variety of ogies like blockchain, are emerging rapidly, facilitating inclu- structured and unstructured data becoming available from sion. These could ease SME access to digital payments, loans, different sources, including social media. Organizational supply chains, land titles, contracts, or even ID. challenges also exist, such as internal resistance to adopting data analytics as a new way of doing business (Bain and Policies for SMEs in the data-driven economy Company 2013). On the other hand, wider access to differ- In an interconnected world, access to and use of digital tech- ent tools can help SMEs “turn digital” and help mitigate the nologies and data tools has become key to SME competi- challenges SMEs face. tiveness, affecting the very chances to survive and develop. Fifth, infrastructure constrains many SMEs in emerging Cloud computing, in particular, allows smaller firms to markets because of challenges in accessibility, affordability, overcome the barriers associated with the high fixed costs and quality of connectivity, particularly outside major of ICT investment, and can help smaller firms rapidly scale urban centers (map 5.1). SMEs scattered across territor- up, providing high-power computing resources flexibly via a ies, particularly microenterprises and entrepreneurs, face pay-as-you-go model. a digital divide that could hinder the benefits of data SMEs tend to struggle to navigate the web of regulations for SMEs. and policies pertaining to data and understanding the legal Sixth, lack of trust. This is mainly due to the increased and administrative frameworks governing cross-border data digital security risks perceived by potential SME adopters, flows, data protection, data privacy, and personal data, to which is partly also the result of the increasing sophistica- name a few. Data regulatory frameworks are complex, and tion of digital security threats. In addition, the lack of data many SMEs struggle to find the time and resources to fully governance frameworks in many countries, or a lack of comprehend their implications. SMEs may thus limit their awareness of them or an understanding of how to comply, utilization of data. Firms and Data 81 Evidence shows that the lack of appropriate (open) strategy should also implement awareness-raising standards and fear of vendor lock-in, often due to propri- initiatives for SMEs to better understand the value of etary solutions, can be strong barriers to adoption. This is upgrading their technology and fully exploit the potential particularly true for SMEs, which often lack the negotiating of digital data. Such an initiative could include, as appro- power and know-how about advanced ICTs such as cloud priate, capacity-building programs specifically directed computing, data analytics, and the IoT (see OECD 2017). to SMEs. Recent analysis suggests that small firms are often much • Encourage technology adoption and complementary more affected by poorly designed regulatory frameworks investments. It is crucial not only to facilitate the access than large and incumbent firms, limiting their growth and of SMEs to the technology itself, but also to help them reducing overall business dynamism. Policy action to boost make the necessary complementary investments, for the growth prospects of start-ups and SMEs is thus essential. example, in process and product innovation and in ICT The available data also point to systematic differences in the services or in skills. SME engagement with competency adoption of other complex digital technologies across firms. centers or technology diffusion extension services can The following policies can help SMEs benefit from the also be helpful. opportunities of the data-driven economy: • Implement data security strategies, with SMEs as a specific segment. Data security strategies often look just at the critical • Implement a national digital transformation strategy for information infrastructure, but they should also address the SMEs. Enhancing competition in broadband internet to specific needs of SMEs by providing them with practical increase speed and reduce costs, promoting nationwide guidance and the appropriate incentives for adopting good cloud service markets, or reducing import duties and practices (see EU Digital SME Alliance 2017). For example, taxes on information technology equipment are national interest is increasing in tailored standards and certification policies with widespread impact that are particularly schemes developed by or in cooperation with business and likely to benefit SMEs. In addition, specific strategic in leveraging digital risk. choices need to address the needs of SMEs. For example, digital public procurement has caused an increase of • Implement open data for business initiatives. Some open participation of digital SMEs. Awareness and technical government data initiatives focus on transparency and training may be necessary to enable compliance with data accountability and often tend to neglect its economic policies and fully grasp the benefits of big data. A national value. Disproportionate benefits exist from open data Box 5.5 Open data for SMEs: The European Union and Colombia In 2015, the European Union launched the Open Data Incubator for Europe, an incubator for open-data entrepreneurs across Europe that supports the next generation of digital busi- nesses and fast-tracks development of products. Within the six-month incubation program, companies receive up to €100,000 (US$120,000) in equity-free funding, mentoring, business and data training, high-quality media, visibility at international events, and introductions to investors. Over the course of the 20-month project, the incubator has funded 57 companies. Each has contributed to the development of an open-data ecosystem underpinned by economic, social, and environmental benefits. Colombia´s Emprende con Datos is a project that provides support to entrepreneurs through mentoring and advice for the construction of sustainable business models and digital products and services; Colombian entrepreneurs, public entities, and small information and communication tech- nology companies interested in resolving issues of public and social interest can participate in the use of open government data. Support is provided to selected entrepreneurs for 12 to 20 weeks, during which mentors work hand in hand with the entrepreneurs to strengthen their initiatives. Sources: IDC 2017; and Government of Colombia 2018. 82 Information and Communications for Development 2018 to SMEs, sowing the seeds of further growth and inno- therefore been in crafting policy that recognizes data as an vation. Even when government data does not have a price infrastructure asset.11 These government policies typically tag, the availability of data can depend on “who you are focus on management of the data assets (collection, access, and who you know”; often, the relevant official must be reuse, sharing, preservation, security) and data governance persuaded to supply data and sometimes a personal visit (ownership, funding), though some also address storage to the office is necessary. As such, open data democratizes (data localization, data center management). The same access and levels the field with respect to incumbents principles apply to private sector firms gearing up to develop with established relationships and resources. In 2017, the data assets. In addition, governments need to facilitate the World Bank’s Open Data for Business assessment in Kenya development of physical infrastructure to manage data found that small businesses could benefit from the release from nontraditional sources that the current telecom infra- of government procurement, budget, and geospatial data, structure is not designed to support (IoT, for instance, or call and that this would help address structural disadvantages data records). in information access relative to larger, more established, companies. Close the data talent gap The shortage of data skills may be the most serious systemic • Promote data cooperatives among SMEs and value chains. factor holding back data-based innovation and productivity These collaborative pools of data can facilitate access and in several countries. Research suggests that 90 percent of use, and pave the way for moving beyond simply sharing jobs within developed economies already require a measure information to a livelier exchange across public-private of digital and data skills (UN Broadband Commission boundaries. 2017), while less than one-third of the population possesses adequate skills. This is a gap that governments must close Looking ahead quickly. A few good practice examples include the Skills Plus Data inequalities, as noted, increasingly dominate in global program in Norway;12 the Tech Partnership13 (a network economies, but they need not be permanent. Available of employers focused on developing digital skills) and policy options,10 discussed further in chapter 6, include the Doteveryone (an independent think tank focused on the following: digital society) in the United Kingdom;14 the Intel-backed “She Will Connect”15 initiative in Nigeria and Kenya; and the • Developing data infrastructure, though competitive e-schools program in Estonia. market entry • Closing the data talent gap Count on disruption • Anticipating disruption, which may require, for instance, The current wave of digital disruption has produced many more frequent policy reviews and allowing new experi- winners that dominate the economic landscape (described mental approaches to flourish without preemptive by The Economist [2018b] as “Big, Anti-competitive, regulation Addictive and Damaging to Democracy” or BAADD). The disruptors may soon become the disrupted, however, • Further clarifying and improving the policy and regula- especially as even newer types of data sources emerge and tory environment firms with next-wave data skills develop new products and • Promoting data innovation and entrepreneurship services. Other threats include the disruption of the current advertising-based models (Bershidsky 2017), which may Develop data infrastructure suffer if more restrictive data policies become the norm Recognition is growing within many governments that and data ownership is relitigated in different societies. in the digital economy, as an infrastructure asset, data is Others have theorized that decentralized technologies like on par with more traditional infrastructure like trans- blockchain might ultimately be the death knell for firms port and public utilities. Indeed, stock exchanges place a like Google or Facebook (Munster 2017). None of this is much higher value on control of a customer’s data than inevitable and it would be foolish to count the incumbents control of infrastructure (see chapter 1). Recent interest has out, but the age of disruption is not over. Firms and Data 83 Notes Bershidsky, Leonard. 2017. “Google and Facebook Too Can Be Disrupted.” Bloomberg, December 8. https://www.bloomb 1. See https://www.economist.com/news/briefing/21741139 erg.com/view/articles/2017-12-08/google-and-facebook-too -will-be-bad-news-some-global-logistics-business-going-be -can-be-disrupted. -transformed. Boateng, Richard, Joseph Budu, Alfred Sekyere Mbrokoh, Eric 2. See https://ac.els-cdn.com/S2351978917303918/1-s2.0 Ansong, Sheena Lovia Boateng, and Augustus Barnnet -S2351978917303918-main.pdf?_tid=0493ba6c-09e4–11e8 Anderson. 2017. “Digital Enterprises in Africa: A Synthesis of –9ef4–00000aacb360&acdnat=1517773737_248e32a343563 Current Evidence.” http://believeoverhope.org/pearlrichards 316344887ea19da9f17. /index .php/2017/08/20/digital-enterprises-in-africa-a-syn 3. See https://www.mckinsey.com/business-functions/digital thesis-of-current-evidence/. -mckinsey/our-insights/the-case-for-digital-reinvention. Booth, Gregory D. 2017. “A Long Tail in the Digital Age: Music 4. See https://www.mckinsey.com/business-functions/operations Commerce and the Mobile Platform in India.” Asia Music /our-insights/digital-transformation-raising-supply-chain 48 (1): 85–113. -performance-to-new-levels. Brahm, Chris, Aaron Cheris, and Lori Sherer. 2016. “What Big 5. For more information, see https://www.capterra.com/customer Data Means for Customer Loyalty.” Brief, Bain and Company, -service-software/. August 7. http://www.bain.com/publications/articles/what 6. See http://www.m-kopa.com/. -big-data-means-for-customer-loyalty.aspx. 7. See http://offgrid-electric.com/. Brynjolfsson, Erik, Lorin M. Hitt, and Heekyung Hellen Kim. 8. See https://www.pegafrica.com/. 2011. “Strength in Numbers: How Does Data-Driven Deci- sion Making Affect Firm Performance?” Unpublished. https:// 9. See http://www.bboxx.co.uk/. papers.ssrn.com/sol3/papers.cfm?abstract_id=1819486. 10. For more information, see World Bank 2017. Burson, Forrest. 2016. “6 Ways Amazon Is Changing Supply Chain 11. See https://theodi.org/what-is-data-infrastructure. Management in 2016.” Software Advice. https://www.software 12. See https://www.kompetansenorge.no/English/Basic-skills advice.com/resources/amazon-supply-chain-management/. /Competenceplus/. BusinessTech. 2018. “A First Look at Amazon’s New Offices in Cape 13. See https://www.thetechpartnership.com/. Town.” https://businesstech.co.za/news/business/227745/a 14. See https://doteveryone.org.uk/. -first-look-at-amazons-new-offices-in-cape-town/. 15. For more information, see Intel 2013. Carey, Bjorn. 2016. “Smartphone Speech Recognition Can Write Text Messages Three Times Faster Than Human Typing.” Stanford News Services, August 24. https://news.stanford.edu/2016/08/24 References /stanford-study-speech-recognition-faster-texting/. Andrews, D., C. Criscuolo, and P. N. Gal. 2016. “The Best versus Casselman, Ben. 2017. “A Start-Up Slump Is a Drag on the the Rest: The Global Productivity Slowdown, Divergence Economy; Big Business May Be to Blame.” New York Times, across Firms and the Role of Public Policy.” Productivity September 20. https://www.nytimes.com/2017/09/20/busi Working Paper 5, OECD Publishing, Paris. https://www.oecd ness/economy/startup-business.html. .org/global-forum-productivity/library/OECD%20Produc Constantinou, Ioanna, Attila Marton, and Virpi Kristiina Tuun- tivity%20Working%20Paper%20N%C2%B05.pdf. ainen. 2017. “Four Models of Sharing Economy Platforms.” Armstrong, Mark. 2006. “Competition in Two-Sided Markets.” MIS Quarterly Executive 16 (4): 231–51. http://misqe.org RAND Journal of Economics 17 (3): 688–91. /ojs2/index.php/misqe/article/viewFile/798/474. Badran, Mona Farid. 2014. “Access and Use of ICTs in Female- da Silva, Issa Sikiti. 2014. “Mali Digital Plan 2020 to Reorganise Owned SMEs in Selected Arab Countries and Brazil: A Economy.” Biztech Africa, December 9. http://www.biztech Comparative Study.” Working Paper 2014/12, Maastricht africa.com/article/mali-digital-plan-2020-reorganiseecon School of Management. https://www.msm.nl/resources omy/9327/. /uploads/2014/05/MSM-WP2014-12.pdf. The Economist. 2018a. “A Bold Scheme to Dominate Ride-Hailing.” Bain and Company. 2013. “Big Data: The Organizational Challenge.” May 10. https://www.economist.com/briefing/2018/05/10/a https://www.bain.com/insights/big_data_the_organizational -bold-scheme-to-dominate-ride-hailing. _challenge/. ———. 2018b. “The Techlash against Amazon, Facebook and Bakhshi, Hasan, Albert Bravo-Biosca, and Juan Mateos-Garcia. 2014. Google—And What They Can Do.” January 20. https://www “Inside the Datavores: Technical Report.” Nesta. https://www .economist.com/briefing/2018/01/20/the-techlash-against .nesta.org.uk/report/inside-the-datavores-technical-report/. -amazon-facebook-and-google-and-what-they-can-do. 84 Information and Communications for Development 2018 Eisenmann, Thomas R., Geoffrey Parker, and Marshall Van Alstyne. Post, January 16. https://www.washingtonpost.com/business 2011. “Platform Envelopment.” Strategic Management Journal /economy/companies-race-to-gather-a-newly-prized-cur 32 (12): 1270–85. rency-our-body-measurements/2018/01/16/5af28d98-f6e8 Enders, Albrecht, Harald Hungenberg, Hans-Peter Denker, and -11e7-beb6-c8d48830c54d_story.html?noredirect=on&utm Sebastian Mauch. 2008 “The Long Tail of Social Networking: _term=.4eb7f0e18af4. Revenue Models of Social Networking Sites.” European Henten, Anders, and Iwona Windekilde. 2016. “Transaction Costs Management Journal 26 (3): 199–211. and the Sharing Economy.” Info 18 (1): 1–15. Ernst and Young. 2016. “Digital Supply Chain: It Is All About That IDC (International Data Corporation). 2017. “Impact Assessment Data.” https://www.ey.com/Publication/vwLUAssets/Digital of Odine Program: Open Data Incubator Europe.” https:// _supply_chain_-_its_all_about_the_data/%24FILE/EY-digit opendataincubator.eu/. al-supply-chain-its-all-about-that-data-final.pdf. ITU (International Telecommunication Union). 2015. “Interoper- Etherington, Darrell. 2017. “Otonomo Raises $25m to Help Auto- ability in the Digital Ecosystem.” GSR Discussion Paper, ITU, makers Make Money from Connected Cars.” TechCrunch, April 7. Geneva. https://www.itu.int/en/ITU-D/Conferences /GSR https://techcrunch.com/2017/04/07/otonomo-raises-25m /Documents/GSR2015/Discussion_papers_and_Presenta -to-help-automakers-make-money-from-connected-cars/. tions/Discussionpaper_interoperability.pdf. EU Digital SME Alliance. 2017. “European Cybersecurity Strategy: Intel. 2013. “Intel: She Will Connect Initiative.” https://www.intel Fostering the SME Ecosystem.” https://www.cyberwatching.eu .com/content/www/us/en/corporate- responsibility /social /news-events/news/european-cybersecurity-strategy-fostering -impact-and-educational-initiatives/she-will-connect.html. -sme-ecosystem. Kelly, Tim, and Carlo Maria Rossotto. 2012. Broadband Strategies Evans, David S. 2013a. “Economics of Vertical Restraints for Multi- Handbook. Washington, DC: World Bank. https://openknow sided Platforms.” Working Paper, Coase-Sandor Institute for Law ledge.worldbank.org/handle/10986/6009. and Economics, Chicago. https://chicagounbound.uchicago.edu Kenney, M., and J. Zysman. 2016. “The Rise of the Platform /cgi/viewcontent.cgi?article=1187&context=law_and_economics. Economy.” Issues in Science and Technology 32 (3): 61–69. ———. 2013b. “Some Empirical Aspects of Multi-sided Platform Kharpal, Arjun. 2016. “5 Reasons Why Uber Sold Its China Busi- Industries.” Review of Network Economics 2 (3): 191–209. ness to Didi Chuxing.” CNBC, August 1. https://www.cnbc .com/2016/08/01/5-reasons-why-uber-sold-its-china-busi Evans, David S., and Richard Schmalansee. 2016. Matchmakers: ness-to-didi-chuxing.html. The New Economics of Multisided Platforms. Boston: Harvard Business Review Press. Koçoğlu, İpek, Salih Zeki İmamoğlu, Hüseyin İnce, and Halit Keskin. 2011. “The Effect of Supply Chain Integration on Informa- FlashGlobal. 2018. “Walmart: 3 Keys to Successful Supply Chain tion Sharing: Enhancing the Supply Chain Performance.” Management Any Business Can Follow.” April 12. https:// Procedia—Social and Behavioral Sciences 24: 1630–49. https:// flashglobal.com/blog/supply-chain-management-walmart/. www.sciencedirect.com/science/article/pii/S1877042811015448. Frieden, Rob. 2017. “The Internet of Platforms and Two-Sided Lal Das, Prasanna, Stephan Beisswenger, Srikanth Mangalam, and Markets: Legal and Regulatory Implications for Competition Rasit Yuce. 2017. “Internet of Things: The New Government and Consumers.” Working paper, College of Communica- to Business Platform—A Review of Opportunities, Practi- tions, Pennsylvania State University, State College, PA. https:// ces, and Challenges.” World Bank, Washington, DC. http:// ssrn.com/abstract=3051766. documents.worldbank.org/curated/en/610081509689089303 Government of Colombia. 2018. “Emprende con Datos Finalizó /Internet-of-things-the-new-government-to-business-plat con Más 60 Equipos Apoyados” [Start with Data Finished form-a-review-of-opportunities-practices-and-challenges. with More Than 60 Teams Supported]. Ministry of ICT. Mastercard. 2012. “Nigerians Enjoy the Convenience, Pricing http://www.mintic.gov.co/portal/604/w3-article-73884.html. and Choice of Online Shopping—MasterCard Survey.” Press Graef, Inge, Sih Yuliana Wahyuningtyas, and Peggy Valcke. release, May 1. https://www1.mastercard.com/content/intel 2015. “Assessing Data Access Issues in Online Platforms.” ligence/en/research/press-release/2012/Nigerians-enjoy Telecommunications Policy 39 (5): 375–87. -the -convenience-pricing-and-choice-of-online-shopping GSMA. 2017. Embracing the Digital Revolution Policies for Building the .html. Digital Economy. Digital Transformation Report 2017. GSMA. Matthee, Marianne, and André Heymans. 2013. “How South Afri- h t t p s : / / w w w. g s m a . c o m / p u b l i c p o l i c y / w p - c o n t e n t can SMEs Can Become Better Candidates for Export Finance.” /uploads/2017/02/GSMA_DigitalTransformationRep ort2017 Managing Global Transitions 11 (4): 391–407. _Web.pdf. Matthey, James. 2018. “There’s Something Unusual about Simona Harwell, Drew. 2018. “Companies Race to Gather a Newly Halep’s Australian Open Dress.” News.com.au, January 26. Prized Currency: Our Body Measurements.” Washington https://www.news.com.au/sport/tennis/australian-open/there Firms and Data 85 s-something-unusual-about-simona-haleps-australian-open Vol. 3. Dubai: Mohammed Bin Rashid School of Government. -dress/news-story/28d7ce2da03fb00a9ad2bd5a79093156. https://img0bm.b8cdn.com/images / uploads/article_docs McKinsey Global Institute. 2017. “Digital Transformation: Raising /arab_world_online_2017_4ir_d_fadi_salem_34792_EN.pdf. Supply-Chain Performance to New Levels.” https://www.mckinsey Schwab, Klaus. 2017. The Fourth Industrial Revolution. New York: .com/business-functions/operations/our-insights/digital-transform Crown Business. ation-raising-supply-chain-performance-to-new-levels. Sherman, Alex. 2018. “Uber Is Preparing to Sell Southeast Asia Unit ———. 2018. “Digital Reinvention: Unlocking the ‘How.’” https:// to Grab in Exchange for Stake in Company: Sources.” CNBC www.mckinsey.com/~/media/McKinsey/Business%20Func News, February 16. https://www.cnbc.com/2018/02/16/uber tions/McKinsey%20Digital/Our%20Insights/Digital%20 -preparing-to-sell-southeast-asia-unit-to-grab.html. Reinvention%20Unlocking%20the%20how/Digital-Reinven Still, Kaisa, Marko Seppänen, Heidi Korhonen, Katri Valkokari, tion_Unlocking-the-how.ashx. Arho Suominen, and Miika Kumpulainen. 2017. “Business Munster, Brett. 2017. “Could Blockchain Disrupt Facebook and Model Innovation of Startups Developing Multisided Digital Google’s Business Model?” Medium. https://medium.com Platforms.” In 2017 IEEE 19th Conference on Business Informat- / road-less-ventured/could-blockchain-disrupt-facebook ics (CBI 2017), Vol. 2, 70–75. http://ieeexplore.ieee.org/stamp -and-googles-business-model-fda614de492d. /stamp.jsp?tp=&arnumber=8012942&isnumber=8012382. OECD (Organisation for Economic Co-operation and Develop- Suominen, Kati. 2017. “Accelerating Digital Trade in Latin America ment). 2017. “Key Issues for Digital Transformation in the and the Caribbean.” Working paper, Inter-American Develop- G20.” G20 Germany Presidency and OECD Conference, ment Bank, Washington, DC. https://publications.iadb.org January 12. https://www.oecd.org/g20/key-issues-for-digital /bitstream/handle/11319/8166/Accelerating-Digital-Trade -transformation-in-the-g20.pdf. -in-Latin-America-and-the-Caribbean.PDF. Poppe, Krijn J., Sjaak Wolfert, Cor Verdouw, and Tim Verwaart. Surowiecki, James. 2016. “Why Startups Are Struggling.” 2013. Information and Communications Technologies as Technology Review, June 15. https://www.technologyreview a Driver for Change in Agri-Food Chains. https://s3.ama .com/s/601497/why-startups-are-struggling/. zonaws .com/academia.edu.documents/40043142/542e8 Tan, Barney, Shan Ling Pan, Ziangua Lu, and Lihua Huang. 2015. b a 4 0 c f 2 7 7 d 5 8 e 8 e b f f d . p d f 2 0 1 5 1 1 1 5 - 6 8 2 4 7 - 1 cc j e z q “The Role of IS Capabilities in the Development of Multi-sided .pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A Platforms: The Digital Ecosystem Strategy of Alibaba.com.” &Expires=1537343545&Signature=ZMUDN6eXwMB17wL Journal of the Association for Information Systems 16 (4): 248–80. KmsIDXAZzfR4%3D&response-content- disposition Täuscher, Karl. n.d. “Business Models in the Digital Economy: = inline%3B%20filename%3DInformation_and An Empirical Study of Digital Markets.” Working paper, _Communication_Technology.pdf. Fraunhofer MOEZ, Fraunhofer Center for International Priyadarshinee, Pragati, Rakesh Raut, Manoj Jha, and Bhaskar B. Management and Knowledge Economy, Leipzig. https:// Gardas. 2017. “Understanding and Predicting the Determin- www.imw.fraunhofer.de/content/dam/moez/de/documents ants of Cloud Computing Adoption: A Two Staged Hybrid /Working_Paper/Working_Paper_Digital_Marketplaces SEM-Neural Networks Approach.” Computers in Human _final.pdf. Behavior 76: 341–62. Täuscher, Karl, and Sven Laudien. 2018. “Understanding Platform Ragoobar, Tricia, Jason Whalley, and David Harle. 2011. “Public Business Models: A Mixed Methods Study of Marketplaces.” and Private Intervention for Next-Generation Access European Management Journal 36 (3): 319–29. Deployment: Possibilities for Three European Countries.” TechMoran. 2017. ”Jumia Launches Offline Experiential Centers in Telecommunications Policy 35 (9): 827–41. Kenya.” September 14. https://techmoran.com/jumia-launches Rochet, Jean-Charles, and Jean Tirole. 2003. “Platform Competi- -offline-experiential-centers-in-kenya/. tion in Two-Sided Markets.” European Economic Association Tsai, Wen-Chun. 2016. “Analyzing the Emergence of Alibaba 1 (4): 990–1029. Group from Business Ecosystem Perspective.” Journal of Rohlfs, J. 1974. “Theory of Interdependent Demand for a International Management Studies 11 (2): 53–64. Communications Service.” Bell Journal of Economics and UN Broadband Commission. 2017. “Working Group on Educa- Management Science 5 (1): 16–37. tion: Digital Skills for Life and Work.” Report, Broadband Ruutu, Sampsa, Thomas Casey, and Ville Kotivirta. 2017. “Develop- Commission for Sustainable Development. http://unesdoc ment and Competition of Digital Service Platforms: A System .unesco.org/images/0025/002590/259013e.pdf. Dynamics Approach.” Technological Forecasting and Social UNCTAD (United Nations Conference on Trade and Development). Change 117: 119–30. 2017. Information Economy Report 2017: Digitalization, Trade Salem, Fadi. 2017. The Arab World Online: Digital Transformations and Development. Geneva: UNCTAD. http://unctad.org/en and Societal Trends in the Age of the 4th Industrial Revolution. /PublicationsLibrary/ier2017_en.pdf. 86 Information and Communications for Development 2018 Van Alstyne, M. 2016. “Platform Shift: How New Biz Models ———. 2017. “Internet of Things: The New Government to Are Changing the Shape of Industry.” Video. https://www Business Platform—A Review of Opportunities, Practices, and .youtube.com/watch?v=8OFRD66pI0Y. Challenges.” Washington, DC. http://documents.worldbank World Bank. 2016. World Development Report 2016: Digital .org/curated / en/610081509689089303/Internet -of-things Dividends. Washington, DC: World Bank. doi:10.1596/978- -the-new-gov ernment-to-business-platform-a -review-of 1-4648-0671-1. -opportunities-practices-and-challenges. Firms and Data 87 Chapter 6 Policies for the Data Economy Introduction Second, the chapter considers policies geared toward build- ing consumer trust and principles for setting limits on what A s seen throughout this report, data is at the can be done with data, such as data protection and privacy heart of the digital economy—it is the raw (including cross-border data flows and data localization) and material for the development of new products data security. Trust encourages governments, the private sector, and services and refinement of existing ones. It is also a new and users to innovate and benefit from the data revolution. asset class, worth billions of dollars. Data is also a policy Third, data security is examined. The lack of a secure area that has been evolving very rapidly in the past few years and trusted environment could delay both the adoption of because of the rapid changes in technologies and their effects data-enabled services and products and, potentially, their on relevant data policies. Policy makers struggle to keep offering by the private sector. This could put emerging up, despite efforts to issue technology-neutral data policy market economies at a disadvantage in participation in frameworks. As nontraditional sources of data become more global innovation, educational, and commercial networks. common, and data is used in entirely novel ways, questions Finally, the chapter covers complementary policies that arise about who owns what data, who can do what with it, facilitate building a data economy. Those include policies to and what protections are afforded to whom. support innovation, and those that help build digital skills One overarching message is that data policies can achieve and entrepreneurship. greater impact using a dynamic ecosystem approach. As a rapidly evolving area of policy, the examples Governments need to play a multidimensional role and presented here focus on recent changes. Many come from create new partnerships with a wide range of stakeholders the European Union (EU), as the European Commission to achieve them. This chapter discusses four dimensions of (EC) is moving faster and farther than most in this area. this question. The potential ease of transferring data, including through First, the chapter briefly reviews policies for building accessing websites across borders, gives the standards an strong data infrastructure to support making data avail- international dimension (figure 6.1). Thus, new regulations able, including those for management of data assets and such as the General Data Protection Regulation (GDPR) are data governance. It focuses on open data and principles of interest for the principles they espouse and for the prac- for data sharing between government, businesses, and tical issues they raise for anyone digitally interacting with individuals. them, including developing countries. 89 Figure 6.1 A framework for data policies Reuse of public Sharing private sector data data Privacy Freedom of Data protection information Data policies for building trust Consumer Intellectual Collection Access Reuse Sharing Preservation Cybersecurity protection property rights Data infrastructure Data security Competition Cross-border Data localization data flows Data skills Entrepreneurship Data innovation Policies for building data as an returns to encourage data sharing (OECD 2015). This calls for infrastructure asset ensuring a relevant legal framework exists, with policies aiming for the extensive sharing, use, and development of public data Data as infrastructure has recently become a prominent sources and research data infrastructure. Policies governing topic for discussion in policy circles. Recognition is grow- business-to-business and business-to-government data also ing within many governments that, in the digital economy, need to encourage appropriate sharing of data, spurring data is a critical new infrastructure asset that enables new, innovation while avoiding stifling competition. often more efficient and inclusive delivery of other activ- Governments increasingly recognize government data as ities, particularly services. The value of the data and the a strategic resource (the data management policy of Qatar’s potential for use expands with quantity and quality. Interest Ministry of Information and Communications Technology has surged in crafting policies to support wider use of data, [2015], for instance, explicitly identifies it as such). The while also recognizing the new risks and challenges it poses. New Zealand data and information management principles This first section looks at policies to expand the sharing of provide a useful set of principles stating that information data and the next looks at how to balance that with address- should be open, readily available, well managed, reasonably ing concerns about privacy and security. priced, and reusable, unless there are necessary reasons for Greater access to data also has beneficial spillovers, and its protection (Government of New Zealand 2011). Personal data can be used and reused to open up significant growth and classified information will remain protected. Govern- opportunities or to generate benefits across society in ways ment data and information should also be trusted and unforeseen when the data was created. The Organisation for authoritative.1 Economic Co-operation and Development (OECD) therefore In supporting strong data infrastructure, governments recommends that policy makers aim for an innovation policy should consider policies focused both on management mix that encourages investments in data (its collection, cura- of data assets and on data governance. The next section tion, and reuse), while addressing the low appropriation of 90 Information and Communications for Development 2018 considers these issues for three types of data sharing—the this data suboptimal for creating value-added services and reuse of certain government data, business-to-business data, products. and business-to-government data. Transparency has been an objective of many open data initiatives in the past decade, based on the principle that Policies on the reuse of public sector information sunlight is the best disinfectant.2 The most important Many now recognize public data in user-friendly formats data sets that help enable the growth of an anticorruption freely available online for anyone to use and for any purpose culture have now become clearer: corporate registers, public as a major resource to aid economic growth. While social contracting information, information on public officials, media, companies, and nongovernment organizations can land registration information, government budget and all be sources of open data, the term is usually applied spending data, and courts data are all helpful for this to data that comes from government and government- agenda. Moreover, the modalities of the publication of supported institutions—open government data. The data this data are also important. Data published in user- governments collect or generate, when freely available, is friendly, machine-readable formats helps governments fight more than just a tool to hold governments accountable. It corruption more effectively as it enables civil society to also drives innovation that can help launch new businesses, analyze and support government efforts to identify irregu- optimize existing companies’ operations, create jobs, and larities. Promotion of common standards, such as the Open improve the climate for foreign investment (World Bank Contracting Data Standard, enables sharing of toolsets so 2014). Increased availability of data can fuel the private that local activists can build on the work of those in other sector through access to new types of public and publicly jurisdictions. funded data, including data held by utility companies and Although anticorruption has been an objective of many the transport sector, and research data. open data initiatives in the past decade, the supply of data The benefits connected to reusable public data, espe- alone is seldom sufficient3—civil society actors who will use cially open government data, are diverse and yet largely the data, as well as government willingness to respond, are untapped. Positive outcomes range from greater transpar- needed; with these in place, a “virtuous circle” can be created ency, efficiency, and economic growth to broader social in which some initial pressure leads to initial improvement welfare. Although some countries are applying the “open- and release of more data. by-default” principle to public data sharing policies, particu- Data access policies are increasingly expanding to cover larly advanced economies, this does not imply that all data data generated by publicly funded research. “Open science” sets should be made available to the public. When thinking efforts rely on the premise that scientific information of open data policies, governments can consider that the resulting from public funding should be accessible and same limits apply to open data as to access to information. reusable, with as few restrictions as possible. The opening In other words, the protection of privacy, personal data, of research processes, designs, workflows of dissemination or national security are common limits. In addition, for of results, and methodologies can expand quality, avoid governments just beginning to open their data, opening duplication, and facilitate reuse, which ultimately can help certain data sets over others holds more potential value. maximize the societal role of science. Research data policies Geospatial data, or data on weather, transport, and roads, need to ensure coherence and complementarity between can be particularly critical, and among the first any govern- open access and open data policies. ment should consider opening. Governments can also promote business opportunities by The Open Data Charter sets out six principles developed mainstreaming the use of application programming inter- in 2015 by governments, civil society, and experts around faces (APIs) for more automatic access to dynamic data. This the world to represent a globally agreed-on set of aspi- has important implications in supporting data ecosystems, rational norms for how to publish data (table 6.1). So far, as it saves costs and time to access data, facilitating prac- 57 national and local governments have adopted it for the tical usage. Sharing data through secure APIs can produce development of open data policies. However, a vast amount value added for data assets across the data value chain, of public information is still made available (if at all) in particularly where potential is often unexploited by data non-user-friendly formats (that is PDFs and JPEG), making holders. However, current public sector use of APIs is limited, Policies for the Data Economy 91 Table 6.1 Open data principles 1. Open by default This can represent a real shift in how governments operate and how they interact with citizens. The presumption is that governments need to justify data that is kept closed, for example, for security or data protection. To make this work, citizens must also feel confident that open data will not compromise their right to privacy. 2. Timely and comprehensive Open data is only valuable if it is still relevant. Getting information published quickly and in a comprehensive way is central to its potential for success. As much as possible, governments should provide data in its original, unmodified form. Maintaining historical data is important for keeping track of changes and evaluating the impact of reforms. 3. Accessible and usable Ensuring data is machine readable and easy to find will make it go further. Portals are one way of achieving this. But it is also important to think about the user experience of those accessing data, including the file formats in which information is provided. Data should be free of charge, under an open license, for example, those developed by Creative Commons. 4. Comparable and interoperable Data has a multiplier effect. The more quality data sets you have access to, and the easier it is for them to talk to each other, the more potential value you can get from them. Commonly agreed-upon data standards play a crucial role in making this happen. 5. For improved governance and citizen engagement Open data has the capacity to let citizens (and others in government) have a better idea of what officials and politicians are doing. This transparency can improve public services and help hold governments to account. 6. For inclusive development and innovation Finally, open data can help spur inclusive economic development. For example, greater access to data can make farming more efficient or it can be used to tackle climate change. Finally, we often think of open data as just about improving government performance, but a whole universe exists of entrepreneurs making money from open data. Source: Open Data Charter. Box 6.1 Defining a policy framework for open data: Mexico’s experience In 2015, Mexico aimed to make government public data available to all citizens in user-friendly formats on the data.gob.mx platform. In 2013, the Open Data Readiness Assessment was conducted, laying the foundations for implementing the country’s open data initiative. The steps taken resulted in (a) the implementation of a national Open Data policy; (b) the estab- lishment of the Consultative Council composed of representatives from the private sector, civil society organizations, and academia; (c) the launch of the single data catalog; (d) imple- mentation of programs for data use in the elaboration of public policies; (e) identification and implementation of the reuse sector; and (f) creation of the Data Squad for preparation and publication of data among public officials. With these measures, Mexico ranks first among the Latin American and the Caribbean countries in three out of four of the Open Data Barometer’s evaluations of the country’s preparedness for open data. particularly in developing countries—expanding use requires This can allow companies to connect in different data- awareness raising and training. sharing engagements with larger companies, small and medium enterprises (SMEs), and start-ups, or even the Policies for private sector data as a driver of innova- public sector. This way, data value can be maximized on tion and competitiveness several fronts simultaneously. Data can be shared to support the creation of more Data-sharing models have emerged to promote fair than one new product, service, or production process. and competitive markets for products and services that 92 Information and Communications for Development 2018 rely on nonpersonal machine-generated data created, and (b) respect for the commercial interests of data holders to assist public agencies in accessing private sector data, and users; (c) ensuring undistorted competition when to guide policy decisions or improve public services. The sharing sensitive data; and (d) minimizing data lock-in to EC (2018) defines a set of key principles to be taken into enable data portability as much as possible. account to improve data sharing for all parties involved, in • Promoting the development of trusted and secure plat- business-to-business (B2B) and business-to-government forms and privacy-minded analytical techniques to secure situations. Access to and reuse of private sector data also sharing of proprietary industrial data and personal data constitute major cornerstones of a common data economy. and ensure compliance with relevant legislation (data Business-to-business data sharing protection, IP rights, and so on). Data collaboratives An ever-increasing amount of data is created automatically have emerged as a potentially viable response to the by objects or processes based on disruptive technologies, data challenges companies face. They provide access such as sensors and the Internet of Things. These mainly to “verified” and useful data (open data or otherwise) relate to nonpersonal data generated by machines and open from public and private sources, commercial models a new discussion and a dilemma around the privileged that reward data producers and consumers, legal and position of the producers of those devices in determining regulatory protections and guidance, data security infra- the access to and usage of the data they generate. structure, network connectivity, analytics infrastructure, An EC public consultation with private sector stakeholders and literacy programs. showed consensus that more B2B data sharing would be bene- ficial (EC 2018), where data can be reused without losing data Business-to-government data sharing quality or competitive advantage. The critical point in B2B data Data that companies collect and produce—cellular data, sharing might not rely on ownership, but on how data access is utility companies, shared carpooling services (such as Uber), structured, managed, and approached. It could be argued that, or social media—can lead to improved traffic, better urban at this initial stage of the development of data economies, it is planning, and so on. As with B2B data sharing, governments too early for legislation requiring B2B data sharing. However, can consider using key principles to guide these exchanges. governments can consider nonregulatory measures to promote The EC has defined the following key principles: (a) propor- B2B data sharing: tionality in the use of private sector data justified by clear and demonstrable public interest—the cost and effort • Fostering the adoption and use of APIs for easier and required for the supply and reuse of private sector data more systematic access to data. APIs can open up a data should be reasonable compared with the expected public ecosystem of startups, exploiting unused data sets and benefits; (b) purpose limitation of business-to-government supporting host organizations to adopt and create new collaboration; (c) “do no harm”—protection of trade secrets data services and products. This has happened in the and other commercially sensitive information; (d) condi- financial sector, leading to the emergence of financial tions for data reuse; and (e) mitigate limitations of private technology ecosystems and new products and services sector data such as potential inherent bias—companies that are already showing a relevant impact on banking supplying the data should offer reasonable and proportion- the unbanked. The configuration and utilization of APIs ate support to help assess the quality of the data. requires the consideration of several principles: security, use of standards, user-friendliness, stability, and sustain- ability over time. Data policies for building trust • Providing key guiding principles for good practices in What is at stake? B2B sharing agreements to ensure fair and competitive Policies ruling the data governance framework require as markets and to avoid excluding SMEs. Those crafted much attention as is given to the need for robust manage- by the EC are an example, including (a) transparency, ment of data infrastructure. This section focuses on data clearly identifying who will have access, to what type protection and privacy, as well as data security policies. of data, at which level of detail, and usage purposes; Countries are struggling with how to build trust in the Policies for the Data Economy 93 digital data economy. Policy considerations in many cases individual who ended or began journeys at a known lesbian, are similar to those posed by “analog data.” Personal data, gay, bisexual, and transgender (LGBT) destination. In coun- whether machine generated or not, is subject to the same tries where this is condemned by law, this information could privacy rules in Europe as analog, for instance. The World result in people being sent to prison or worse. Intellectual Property Organization argues that no addi- The struggle between the need to protect privacy and tional intellectual property protection should be awarded to allowing big data to continue to improve the way we live machine-generated data beyond the traditional ones (Burns without quashing innovation is unlikely to be resolved 2017), nor should it be awarded less. A recent (2018) paper easily. It remains to be seen how effective new laws, such by the World Bank and the Consultative Group to Assist the as the GDPR, will be in achieving either of these aims. Poor (CGAP), looking at the use of alternative data to build One certainty is that data production will not slow down credit histories for greater financial inclusion, frames many and neither will the development of new ways to use policy questions that, at their core, do not vary much from it—legislators face an uphill battle to keep pace. those posed by the use of other kinds of data. If countries can tackle key aspects of trust policies, they will be well on Trends in principles in protecting data privacy their way to creating a better environment for the digital Data protection laws have been in place for a while, with data economy. privacy rights protected as early as in ancient Roman times. Data protection and privacy are critical policy issues for Even though the right to privacy was not recognized as such the data economy and are key to building consumer trust. by Roman law, several privacy violations, such as the inva- These are two separate but intertwined concepts. Data sion of the sanctity of one’s home, were covered under the protection refers generally to the protection of personal data, law. However, the numerous issues brought up by the sheer though it may also be used in the context of commercially volume of data that can be collected about a person through sensitive data. A common definition of personal data was online personas and questions about who can do what with cemented by the EU in its 95/46 Directive,4 as data that those data, or who “owns” that data, have brought privacy identifies a person, or allows such identification by cross- concerns to the forefront of the news and therefore to the referencing it with other available data. Privacy is a broader desks of policy makers. concept, which has sometimes been defined as the right to Technological developments are pushing policy makers be let alone (Warren and Brandeis 1890), and it refers not to either amend existing privacy legislation or pass new only to data, although it partially covers it. What is at stake legislation. The EU 95/46 Directive was replaced in May for either goes beyond keeping personal or embarrassing 2018 by the GDPR. Even international standards, such as information from others. Big issues are how the data will be the OECD Privacy Guidelines (OECD 1980, Rev. 2013), or used and the risks that it will exclude people (such as, for the APEC Privacy Framework (APEC 2015), were recently example, by making people ineligible for insurance or credit) revised or expanded to accommodate these developments. or be used for price discrimination, to suppress competition, The GDPR is likely to have a trickle-down effect as other or to manipulate people (such as, for example, through the countries revise data protection legislation, and it is already crafting of news that could swing elections). This chapter having extraterritorial reach in private sector behavior. uses “data protection” and “privacy” interchangeably. Consumers around the world are getting notices of revised Big data provides an example of the massive challenge to privacy policies by global companies in compliance with privacy in widespread use of technology. Big data preserves the GDPR, and some content websites outside Europe have privacy by detaching information from individuals and refused access to European consumers because they could repurposing it. However, by taking multiple, anonymized not ensure compliance with the GDPR (Noack 2018). data sets and triangulating them, you can begin to break Among the regulatory trends in the privacy space, the down that anonymity. For instance, take information about more salient are focused on a risk-based approach to compli- all the journeys that people have taken over the past year ance and on proactive measures to protect privacy, as opposed from a taxi service. This data alone is not necessarily sensi- to measures in reaction to a breach. “Privacy-by-design” stan- tive, but if you combine it with venue information and social dards require companies to embed technological measures media, you could conceivably make assumptions about an protecting privacy in their product and service design, 94 Information and Communications for Development 2018 for instance, to ensure anonymity of users. The concept of that are data processors or controllers and found not to be privacy by design was developed by Ann Cavoukian, former compliant. Firms in emerging markets may be subject to those Information and Privacy Commissioner in the Canadian fines if they are found not compliant. This could happen, for Province of Ontario. instance, to firms with no or little physical presence in the EU, The GDPR takes this one step further by carving out but whose advertisements target EU consumers. Enforcement the related concept of “privacy by default.” With privacy by could be done through the branch office or subsidiary located default, the expectation is that companies and those process- within the EU of the firm from the emerging market. ing or controlling the data will put in place mechanisms to ensure that only those items of personal information needed Issues to consider when enacting or updating a for each specific purpose are processed “by default.” The main legal framework principle is to be proactive rather than reactive and preventive Follow the principles instead of remedial. This is a practical approach for emerging Numerous countries have identified the need for coordination markets to consider, since it can help enforcement, give the and cooperation in privacy and data protection. Most have private sector a more proactive role, and help prevent privacy used regional bodies, such as the EU or Asia-Pacific Economic violations and data breaches. The capabilities of the local Cooperation, as the vehicle for that cooperation rather than private sector would need to be considered, as well as a plan international agreements. These bodies have enacted guide- that would help them ease into this approach should they not lines or regulations and, not surprisingly, many principles have the necessary resources or skills to do so. are common among them. Countries considering enacting Another trend is a strong focus on data security, prevent- or updating privacy legal frameworks can reference those ive as well as once a data breach has occurred, on breach common principles. These include principles shared by the notifications. This important corollary to privacy protec- European Convention on Human Rights,5 the OECD Privacy tions is discussed in more detail below. Guidelines, the APEC Privacy Framework, and the GDPR. A legal framework is increasingly a necessity, Raise awareness, highlight key issues, engage not a luxury relevant stakeholders early on The need for a legal privacy framework is no longer At least 35 countries are currently drafting data protec- questioned, even in emerging markets that perhaps are tion laws to address this gap (UNCTAD 2018; map ES.1). used to seeing privacy at some point as a “luxury” right. A number of economies are also considering reforms to According to the United Nations Conference on Trade and legal frameworks, including to the extent that it may be Development (UNCTAD 2018), 108 countries either had affected by extraterritorial application of the EU’s GDPR. data protection laws or some kind of law that deals with However, drafting and implementing data protection laws data, whether in force or not, as of April 1, 2018. However, is time consuming and challenging. Surveys by UNCTAD of levels of protection, particularly enforcement, vary widely, government representatives in 48 countries in Africa, Asia, even within countries with legal frameworks. In the nearly and Latin America and the Caribbean (UNCTAD 2016) 30 percent of countries with no laws in place, personal data point to the need to build awareness and knowledge among receives little or no protection, reducing trust and confi- lawmakers and the judiciary to formulate informed policies dence in a wide range of commercial activities. A lack of or and laws in data protection and to enforce them effectively. weak regulation put these countries at risk of being cut off More than 60 percent of respondents reported difficulty from international trade opportunities, because many trade understanding legal issues related to data protection and transactions require cross-border data transfers that comply privacy. Similarly, 43 percent noted a lack of understand- with minimum legal requirements. ing among parliamentarians and 47 percent among police The GDPR brings additional incentive to strengthen data or law-enforcement bodies, which can delay adoption and protection regimes for countries without one or with weak enforcement of data protection laws. regimes. With a clear objective for extraterritoriality beyond Another study, covering 22 of the globe’s largest infor- EU borders, the GDPR introduces fines of up to €20 million, mation and communication technology (ICT) firms, found or 4 percent of global turnover, whichever is higher, for firms that none of them disclosed adequate information about Policies for the Data Economy 95 privacy practices and how user information is collected, is conducting stakeholder consultations. This will help its shared, retained, and used.6 Requests by governments for government put together a framework taking into account access to such data are growing, with most emanating from potential obstacles for implementation, including those that the United States (figure 6.2). could come from a lack of private sector capacity or other As with any other legal issue, when developing a legal country-specific issues. framework, it is important to engage the main stakeholders early on. This will allow countries to understand potential Consider the legal culture issues and to get their buy-in and build capacity for imple- As seen in chapter 4 and figure ES.4, the level of tolerance mentation. Some countries, like Mexico, have put together a for giving up one’s privacy varies from country to country, comprehensive effort to raise awareness of these issues with and from individual to individual, with some citizens putting the different government branches, including at the state greater value on protecting their credit card information and level. With the support of a World Bank project (World Bank others placing greater value on protecting their health infor- 2008), the Ministry of Economy in Mexico commissioned mation. For any country considering enacting a new privacy a thorough review of existing legal issues and gaps and put framework, this cultural dimension will be critical. A reflection together a training package for the judiciary, parliamentarians, of these cultural issues can be seen in the different approaches and state government officials. This helped Mexico not only to to privacy taken by the EU and the United States. In the EU, identify issues, but also to prepare for implementation. the right to privacy, but also the right to have one’s personal Other countries, particularly in Eastern Europe, have data protected, are considered fundamental and are recognized benefited from twinning programs with an existing data in the Charter of Fundamental Rights of the European Union protection agency in another country that has provided (2012, Articles 7 and 8). This approach has resulted in the technical assistance in the setting up of both the legal EU having an umbrella data protection framework that does framework and their counterpart agency. It is also impor- not distinguish between data being held by private or public tant to identify potential issues and concerns for private actors, and which contemplates only a few exceptions, such sector stakeholders at an early stage. Countries in the Latin as in the area of national security. But the EU is not alone in American region support each other in such initiatives considering privacy a fundamental right. Even in India, where through the Ibero-American Data Protection Network.7 this was no explicit right to privacy, India’s Supreme Court Seychelles is reviewing its data protection framework and found recently that privacy is a fundamental right protected Figure 6.2 Government requests for user data a. Number of government requests for b. Government requests for Facebook user data, Google user data, 2009–16 second half of 2016 120,000 103,013 100,000 Rest of 87,062 world Brazil 19% 80,000 3% United 66,535 Italy States 60,000 3% 40% 53,356 Germany 42,327 7% 40,000 34,001 27,625 France 7% United 20,000 12,539 India Kingdom 10% 11% 0 2009 2010 2011 2012 2013 2014 2015 2016 Sources: Adapted from Google, Transparency Report; Facebook, Government Requests Report. 96 Information and Communications for Development 2018 by the constitution (Agrawal 2017). In the United States, by Other considerations: Extraterritoriality, trade issues, contrast, privacy is not recognized as a fundamental right. and cross-border data flows Although constitutional limits on the government’s intrusion A key privacy policy issue with potential for big effects on into individuals’ right to privacy can be found in the Fourth a country’s economy is the regulation of cross-border data Amendment, and to some extent in the First and Fifth Amend- flows. McKinsey Global Institute (2016) estimates that flows ments, the right to privacy is more of a balancing act against of data and information now generate more economic value other rights, including very strong rights to free speech and than the global goods trade. Although many of these data freedom of information. This has led to a more segmented flows are concentrated in a handful of large companies, some, approach to privacy protection, with, for instance, a Privacy such as eBay, Amazon, and Alibaba, are really platforms Act for children,8 and for the health sector (Health Insurance allowing SMEs all over the world to becoming mini-exporters, Portability and Accountability Act).9 These deal with data held with an impact across multiple economies. And individuals by government entities and are complemented by different are not being left behind. About 900 million people have pieces of legislation for data held by commercial entities. The used international connections on social media to connect international trend has been toward more comprehensive to networks to find a job, and 360 million take part in cross- privacy frameworks, however, and even in the United States border e-commerce (McKinsey Global Institute 2016). several bills have been introduced for “omnibus privacy laws,” although they have not yet been adopted. Data transfer policies A prime example of the cultural debate is whether privacy Companies need to transfer many different kinds of data laws should protect personal data, even in the absence of across borders in the regular course of business. Those may harm. Most jurisdictions can agree that, when an invasion of include data related to commercial transactions, their own privacy results in financial harm to a consumer, reason exists internal operations, monitoring supply, human resources for government to intervene. Most jurisdictions would also data of global employees, and product support in real time. agree that harm can go beyond financial loss. Injury from loss Countries that regulate data leaving their borders often do so of privacy can take many forms, and can include medical iden- on the basis of privacy and data security concerns. Already in tity theft or “doxing,” which is stalking or extortion coming the 95/46 Directive, the EU regulated data transfers outside of from the dissemination of private facts. The U.S. Federal the EU, and transfers were only allowed to countries the EC Trade Commission (FTC) recently brought a case against determined had adequate data protection. So countries with MyEx.com,10 a porn site that allowed users to seek revenge lower standards risk cutting off opportunities for firms or on their ex by posting their intimate pictures. Victims could individuals in their countries from using platforms, websites, pay to remove intimate images and personal information. But or activities involving the transfer of data with EU countries. beyond that, many people lost jobs or job opportunities, and Governments contemplating restrictions on data transfers were threatened, stalked, and harassed (Ohlhausen 2017). outside their borders based on privacy and data security, In some countries, the use of alternative sources of data above those of existing international standards, may discover (such as credit history downloads from a mobile phone) that global companies decide that it is simpler to block has revolutionized financial inclusion, allowing consumers consumers in a particular jurisdiction from accessing services with little to no credit history to access credit, but could than to try to comply with the data protection rules of that also further raise the bar for others to ever be able to access country. This is the recent example of U.S. advertising tech- credit. However, the use of those same data can lead to price nology companies. Having data policies out of line with larger discrimination on the basis of race or gender, or to denial regional players is a particular problem for small markets in of credit because of data inaccuracy. Questions arise about developing countries that are not part of a wider trade bloc. where to draw the line and whether the basis for government Self-regulation, if it complies with required international action should only be preventing harm. For those who view standards, is another option. Given the reality of global privacy as a fundamental right, there is no need of injury for commerce, even the EC has allowed the use of some self- a government intervention, while the definition of injury regulatory tools to ensure adequate protection of data can be as broad as contravening a person’s expectations with outside of its borders under the 95/46 Directive. Global regard to respect for their privacy. companies could issue binding corporate rules or policies Policies for the Data Economy 97 that are internal to a group of companies and become which have all either proposed or enacted data localization binding once approved by the relevant data protection policies (ITIF 2017). authorities. They could also use model contracts with In Rwanda, the regulatory body, the Rwanda Utilities their subsidiaries that would ensure adequate protection of Regulatory Authority, went a step further and imposed a personal data. The GDPR expands existing mechanisms and fine of US$8.5 million in May 2017 on the mobile operator introduces new tools for international transfers. It offers, MTN for storing customer data in Uganda (Reuters 2017). among other things, adequacy decisions, standard contrac- This is equivalent to about 10 percent of its annual revenue, tual rules, binding corporate rules, certification mechan- and the decision is likely to have a chilling effect on foreign isms, codes of conduct, and so-called derogations. Countries investment into the country, as well as deterring foreign considering these options have to consider enforceability of firms from offering services there. instruments like the codes of conduct, and what happens if Defenders of localization laws cite national security, an institution does not follow its code. protection of personal data, local cultural and historical Beyond self-regulation, another option could be specific context, and economic nationalism as arguments; opponents bilateral agreements between countries, whereby smaller see such laws as a major barrier to trade and competitive- countries offer mutual recognition of rules applied in larger ness. Localization creates its own set of winners and losers trade blocs, such as the EU, similar to the approach used in in the domestic market. It has been argued that localization type approval (that is, homologation) of customer premises laws benefit larger firms at the expense of smaller firms that equipment. The EC signed an agreement on International often do not have in-house data skills and must often pay Safe Harbor Privacy Principles,11 whereby certain companies more per unit of data stored at local firms than might have subject to the FTC’s jurisdiction would commit to protect been available from international data hosting services. data abiding by the same principles spelled out in the On the other hand, localization laws can shelter local directive. This allowed many companies in the United States, firms to develop skills and capacity without the threat of which did not have an “adequacy” finding from the EC, to still competition from international firms. Opponents also cite transfer data to and from Europe. The Safe Harbor agreement issues such as poor data security (Pfeifle 2017) (as many has now been replaced by the Privacy Shield agreement.12 countries with localization laws lack the skills to handle data But even if countries do not impose particular restric- securely) and the risk that localization requirements can tions on data flows, having to comply with different sets of soon become more pervasive and expand to include other data protection rules in different countries becomes costly types of data. for companies, and a de facto trade barrier. This has led some companies to apply the EU standards to their world- Balancing other rights wide operations, as they are considered some of the most Data protection and privacy are not absolute rights and need stringent, hoping to minimize compliance costs. to be balanced against other rights. Some of those rights are access to information, freedom of speech, and the protection Data localization of national or personal security interests. Different countries Data localization policies, in addition to data transfer place different emphasis on different values. For access to policies, affect cross-border data flows, international trade, information, the benefit or the public good of divulging and access to global markets (Chander and Lê 2015). Data certain information needs to be balanced against an indi- localization rules require firms to locate data servers or vidual’s right to privacy. It is generally accepted in many data centers within the borders of a country to store and jurisdictions that individuals with a public persona have a process information. Studies show that data localization lesser expectation of privacy than others. For instance, there and other barriers to data flows impose significant costs: is legitimate public interest in knowing whether a lawyer reducing U.S. GDP by 0.1–0.36 percent; causing prices for who is prosecuting a case of sexual harassment is practicing some cloud services in Brazil and the EU to increase by what they preach (BBC 2018). But even here, public figures 10.5 percentage points to 54 percent; and reducing GDP are increasingly bringing cases involving the violation of growth from 2.4 percent to 1.7 percent in Brazil, China, the privacy and sometimes winning compensation for alleged EU, India, Indonesia, the Republic of Korea, and Vietnam, defamation of character (USA Today 2017). 98 Information and Communications for Development 2018 Implementation issues Enforcers can also have a role in continuous awareness rais- Enforcement ing, both for the data subjects and for those who manipulate As with any laws, when considering privacy laws, enforce- the data. ment is a key issue. Whatever the framework, it is only as worthy as its enforcement. Even absent a specific privacy Data security framework, a strong enforcement agency can still protect On March 22, 2018, the U.S. city of Atlanta was hit by the people’s rights through an interpretation of other existing dreaded SamSam ransomware attack, which brought the laws. This has been the case for the U.S. FTC. The FTC has city’s ICT systems to a halt. Utilities could not collect bill been a strong enforcer of privacy rights by implementing payments, citizens could not pay traffic tickets, the police a more general statue, the Federal Trade Commission Act, had to note complaints by hand—the city’s digital apparatus which gives them jurisdiction to protect consumers from essentially stopped functioning and many departments and deceptive or unfair acts or practices.13 Examples of recent agencies lost several years of data (Wright 2018). Unfortu- cases include privacy cases against Uber, Lenovo, VTech, nately, this was not an isolated incident. In India, a journalist and Venmo.14 was able to obtain unauthorized data from Aadhar,15 the The reverse is also true. Countries with strong legal national digital identification system. And in Mexico, web frameworks on paper that are nonetheless not enforced users were surprised to discover that the voter data of more remain at the same level as countries with no framework than 93 million Mexican citizens was easily accessible online at all. Countries considering institutional arrangements for even though the information was classified as confidential their laws can look at different models of good practices (Chang 2016). around the world. The EU GDPR calls for the establishment It is evident that “not all data is created equal” (Fell and of independent data protection authorities. Following this Barlow 2016). Most data is likely of low value, with more model, some economies have chosen to establish a standalone limited amounts of medium value, let alone high value. And data protection authority, including Canada, Mauritius, and risk varies across types of data assets too. From a business’s South Africa. Other countries have chosen to merge that perspective, it makes little financial sense to spend as much independent authority with the authority protecting access to protecting all assets regardless of their value or the risks information rights, such as Mexico and the United Kingdom. they face. Such a requirement could impose a crippling Other countries, such as the United States, have jurisdiction financial burden. At the same time, most organizations spread among several agencies, with the consumer protec- lack the skills to properly audit their data assets and thus tion agency as one of the main enforcers. Yet another model end up with “orphan assets” littered across systems whose brings the enforcement powers for privacy laws under the value is less than the cost of controls to protect them. What Ministry of Justice, such as in Argentina. organizations should focus on, at a minimum, are their No absolute right or wrong way to think about enforce- “extraordinary assets” critical to them (Kaminski et al. ment exists, as long as the end results are there and the law 2017) as well as data for which costs or breaches of privacy is enforced. Important considerations are identified in the rights could be significant should the data become public. OECD Privacy Guidelines, and include encouraging and If protecting everything equally is not an option, taking supporting self-regulation, whether in the form of codes of this risk management approach should safeguard against conduct or otherwise; providing for reasonable means for deprioritizing sensitive data that is of low financial value to individuals to exercise their rights; providing for adequate a firm that holds it. sanctions and remedies in case of failure to comply with A recent report found significant vulnerabilities in more privacy frameworks; and ensuring no unfair discrimination than three-quarters of applications used by the federal against data subjects (OECD 1980, Rev. 2013). government in the United States (Hesseldahl 2015). Numer- Emerging market economies need to consider what is ous reasons appear to explain this: feasible within their own contexts, taking into account the budget and skills required. Enforcement alliances, both with • Poor data management processes (Virtu 2015)—including other local enforcement agencies, including criminal, as well inconsistent response, unresolved issues, notification as with international enforcement agencies, can help greatly. practices, and lack of data encryption practices Policies for the Data Economy 99 • Legacy systems and old software—still in use in many In a report on data-driven innovation, the OECD (2015) government organizations16 recommends that organizations establish a systematic framework of digital security risk management processes • Poor capacity and skills—due to government’s inability and weld it together with the data value cycle (figure 6.3). to attract top-drawer data-security talent In this framework, the criteria for determining the level • A low priority afforded to security when making tech- of security are based on the acceptable level of risk to the nical infrastructure investments—in a recent study, economic and social activities at stake (OECD 2015) and respondents showed a marked preference for investments not the likelihood of threat. Such an approach is premised in network security and end-point security over invest- on the primacy of data as a socioeconomic asset that justi- ments in data-at-rest security.17 fies the move from a culture of security to a culture of risk management. Data security is not only about ICT; it should also cover “analog” aspects of security (such as vetting of staff and physical access to control buildings). Countries use policies Policies for maximizing the as a tool to manage risks and help respond to actual inci- data economy dents. Data security policies can be in different kinds of laws, In addition to policies for the management and governance including cybersecurity and data protection laws. Govern- of data itself, a number of complementary data-related ments have to consider leading by example and applying policies exist that governments can pursue to support the themselves to strong data security measures, in addition to development of the data ecosystem and ensure that access to what they expect from the ICT industry. opportunities is inclusive. Digital skills and data for innova- The OECD identified the main common principles for tion and entrepreneurship are discussed here. information security in their Guidelines for Information Security and Networks in 2002 and updated them in the Data skills OECD Recommendation on Digital Security and Risk To take advantage of the data economy, more people need management (OECD 2015); these principles were further to have the requisite digital skills. Educational programs spelled out in the Madrid Declaration.18 They emphasize that employ rapid skill training are increasingly demanded risk management, awareness raising, having a prepared- to develop data skills and capabilities for the use of data ness and continuity plan to respond to incidents, and tools for innovators, entrepreneurs, SMEs, other private adoption of security measures to avoid data corruption, sector entities, and government agencies. According to Cisco loss, misuse, or unauthorized access. They also highlight (2015), a shortage of 1 million people to fill data security stakeholder cooperation, including across borders, given jobs will exist over the next five years, and demand for data that most incidents have a multi-country footprint. Robust scientists between 2011 and 2013 alone increased about cybersecurity policies,19 targeting the vulnerability of IT 40 percent. Data skills and tools have become crucial among systems, infrastructure, and networks beyond data, should firms, governments, and, particularly, entrepreneurs. complement data security policies. Different international Data literacy is increasingly considered a core skill, with initiatives have produced or are producing guidelines on some research suggesting that 90 percent of jobs within cybersecurity.20 advanced economies already require a measure of digital For consumers, one of the rising trends is to request data or data skills (EC 2018), while less than one-third of the breach notifications. Breach notifications can be useful to population possesses adequate skills. The gap in developing consumers when their data has been compromised or lost, countries is even wider. This is a gap that governments must since a notification allows them to take corrective action as close quickly. needed. Different countries have different requirements for Governments have employed different models to promote breach notification, and the main differences are the triggers digital literacy. Examples include the following: and timeline for notification. The triggers determine what level of breach is required in order to notify consumers and • Inclusion of digital literacy as part of government-supported can rely, for instance, on the sensitivity of the information basic skills programs, such as the Skills Plus program in accessed and the likelihood that it will be misused. Norway.21 100 Information and Communications for Development 2018 Figure 6.3 Digital security risk management cycle 1. Risk assessment What is the level of risk, that is, the possible effect on objectives? 2. Risk treatment What should I do with the risk, on the basis of my acceptable level of risk Systematic cycle of risk management and in light of the expected benefits? 3. Take 4. Reduce 5. Transfer 6. Avoid to third party End of activity 7. Security measures 8. Preparedness Selection and operation Response, incident management, of the measures and containment plan Source: Adapted from OECD 2015. • Support to advanced digital skills. In the United Kingdom, • The incorporation of coding into school curricula. This is for instance, the Government Digital Services supports done in the e-school program in Estonia27 and similar a range of programs, such as the Tech Partnership22 programs in Denmark, the United Kingdom, and the (a network of employers focused on developing digital United States. skills) and Doteveryone23 (an independent think tank Some lessons and policy recommendations for govern- focused on the digital society). ments to consider from these various digital skills initiatives • Programs aimed specifically at women and girls, who are include ensuring data literacy programs are multistakeholder often underrepresented in the ICT sector. Examples include (including participants from the government, private sector, the Intel-backed She Will Connect initiative in Nigeria, and civil society); building on existing programs, where Kenya, and South Africa,24 and Mozilla Learning’s part- possible, rather than starting from zero; blending traditional nership with UN Women to support a network of web nondigital education with data and digital literacy; bridging literacy clubs in Kenya and South Africa specifically formal and nonformal sources of education, such as using aimed at upskilling girls and women through face-to-face mobile phones as a learning tool in developing countries, peer learning (Dhalla 2016). especially for refugees (UNESCO 2018); and developing societal teaching capacity and mentorship programs. • Mentoring and peer learning based programs. Such programs include Reboot UK (Good Things Foundation 2017), the Swedish IT guide program (which pairs immi- Data innovation grants with elderly Swedes),25 and the “CompiSternli” Companies with huge amounts of data at their disposal program in Switzerland (which pairs children with the and the technical capacity and skilled employees to analyze elderly).26 the data will gain competitive advantage (OECD 2015). Policies for the Data Economy 101 In the digital revolution, access to large and diverse • Private sector innovation. Policies are needed to build data sets is a prerequisite for innovation. Policies geared awareness, capacity, and adoption; and to promote toward unlocking the reuse potential of data can boost cross-cutting uptake for market analysis, financial inclu- the data economy so that businesses and governments sion, value chain integration, and know your customer are not left behind, but put forward at the frontier of across sectors. As discussed in chapter 5, to ensure these innovation. policies reach a majority requires attention to SMEs and Public and publicly funded data can be at the service to the underlying layer that can connect firms to custom- of data-driven innovation. Access and reuse of public and ers, vendors, associations, governments, and so on. publicly funded data can constitute a cornerstone for a data • Citizen-driven innovation. Innovation policy tradition- economy. Policies aiming at making more data available and ally supports the “supply” side by funding research and making data more reusable include policies to lower market development in areas deemed to yield scientific market entry barriers, particularly for SMEs, by reducing charges for results. Demand-driven innovation policies, in which the reuse of public sector information. processes are driven by the end beneficiaries rather than The nature of data-driven innovation also raises new researchers, aim to ensure instead greater relevance and challenges, including how to safeguard competition and uptake. This is the case of data policies that consider to avoid using data as a barrier to the next generation of that social innovation can promote citizen engagement entrants and innovators. Given the value of controlling and creative thinking about alternative ways to provide large amounts of data, there can be winner-take-most services and address problems. An example of this dynamics of companies benefiting from network effects approach in Tanzania, Data Zetu,29 is part of the Data (that is, where the more people that use a platform or Collaboratives for Local Impact program, and aims service, the better the experience of everyone else using it). to empower communities in Tanzania to make better, Although it is beyond the scope of this report to discuss more evidence-based decisions to improve lives. Data competition policies, the treatment of data-sharing poli- Zetu works with stakeholders to build skills and develop cies and the handling of data within intellectual property digital and offline tools that make information accessible rights protections will increasingly be central parts of to everyone. Civic tech, crowdsourced programming, them. Another way that governments can address the risk and open innovation processes to tackle development of excessive first-mover advantage is ensuring that its own challenges can bring together the skills and technology data-sharing arrangements do not result in few re-users needed to make a difference in the lives of those who able to exploit the data in practice. Increased transparency need them most. of public data reuse can allow any company, regardless of size, to be aware of the data available and promote a • Development innovation. Data is also shaping the tradi- broader spectrum of re-users exploiting the social and tional paths of development. The UN-coined “data revo- economic value of data. lution” has triggered novel development approaches that The EU estimates that, in 2016, some 254,850 data help analyze the context, measure impact, and coordinate companies existed across the union, and that the figure project efforts on the ground, among others. Data is a could grow to some 360,000 by 2020 under a high-growth cross-cutting tool for achieving the Sustainable Develop- scenario.28 ment Goals. Development is about knowledge, and data amplifies the power of development assistance as the • Government innovation. Government laboratories such building block of knowledge. as fab labs, data labs, and urban labs have emerged across regions. In 2016, the government of Mexico launched • Data entrepreneurship. Governments should ensure that its Datalab for data analysis to improve Mexico’s public other sources of innovation investment, ICT industry policy formulation and management. Among cities, stimulation, and start-up incubation are playing their Barcelona’s CityLab and Mexico City’s Laboratorio para part in supporting the growth of innovative uses of la Ciudad are examples of municipal level interventions data and of the supporting ecosystem of ICT and other for urban innovation using data. services. In 2015, as noted, the Open Data Incubator for 102 Information and Communications for Development 2018 Europe was launched to support the next generation of 10. Emp Media Inc. (MyEx.com), FTC File No. 162-3052 digital businesses and fast-track the development of their (2018), https://www.ftc.gov/enforcement/casesproceedings /162–3052/emp-media-inc-myexcom. products. Within the six-month incubation program, companies receive up to €100,000 (US$116,000) in 11. The Safe Harbor Agreement was later revoked, following the Schrems case in the ECJ, C-362/14—EUR-Lex—Europa equity-free funding, mentoring, business and data train- EU. ing, high-quality media, visibility at international events, 12. See https://www.privacyshield.gov/welcome. and introductions to investors. 13. Section 5, FTC Act, 15 USC 41–58, as amended. 14. See Uber Tech., Inc., FTC File No. 152–3054 (2017), Policies for data-driven development https://www.ftc.gov/enforcement/cases-proceedings In examining data policies for the digital economy, it is /152-3054/uber-technologies-inc; Lenovo, Inc., FTC File easy to focus on the dark side—combating cybercrime, No. 152–3134 (2017), https://www.ftc.gov/enforcement threats to data security, loss of privacy, and similar matters. /casesproceedings/152–3134/lenovo-inc; VTech Elecs. But the data economy is not only about policies to miti- Ltd., FTC File No. 162–3032 (2018), https://www.ftc gate risks; it is also about policies to maximize value. The .gov / enforcement/cases-proceedings/162–3032/vtech -electronics-limited; PayPal, Inc., FTC File No. 162–3102 true value of data is largely in its use. A strong demand- (2018), https://www.ftc.gov/enforcement/cases-proceedings side “pull” of data is important. It creates and maintains /162–3102/paypal-inc-matter. pressure on expanding ubiquity. And it ensures that the 15. https://www.livemint.com/Opinion/MUPJK28VMeoICzl wider data ecosystem develops and that data is turned into 1whSBrJ/Clearing-the-air-on-Aadhaar-data-breach.html. economic or social value with positive impacts for citizens. 16. See https://www.gao.gov/assets/670/669810.pdf. As shown throughout this report, the “pull” can come from 17. See https://dtr.thalesesecurity.com/2017/pdf/2017-thales governments, civil society organizations, the private sector, -data-threat-report-brazil-edition-pr.pdf. academia, journalists, international organizations, and 18. See http://www.privacyconference2009.org/media/notas donors, as well as from individual citizens. Data-driven _prensa/common/pdfs/061109_estandares_internacionales development involves all of us. _en.pdf. 19. For the definition of data and information security versus the definition of cybersecurity, see ttp://nvlpubs.nist.gov Notes /nistpubs/ir/2013/NIST.IR.7298r2.pdf. 20. For instance, see https://ccdcoe.org/multimedia/national 1. The Open Data Initiative provides more details at https:// -cyber-security-strategy-guidelines.html and https://apec theodi.org/topic/data-infrastructure/. .org/Groups/SOM-Steering-Committee-on-Economic 2. Quote attributed to Justice Louis D. Brandeis, available at -and-Technical-Cooperation/Working-Groups/Telecom https://www.brandeis.edu/legacyfund/bio.html. munications-and-Information. 3. See, for instance, Lindstedt and Naurin 2005 (2), which notes: 21. See https://www.kompetansenorge.no/English/Basic-skills “Taken one at a time transparency and free and fair elections /Competenceplus/. will not help much to reduce corruption. Taken together, on 22. See https://www.thetechpartnership.com/. the other hand, they can be a powerful team.” 23. See https://doteveryone.org.uk/. 4. See Directive 95/46/EC on the protection of personal data at https://eur-lex.europa.eu/legal-content/EN 24. See https://www.intel.com/content/www/us/en/corporate /TXT/?uri=CELEX:31995L0046. -responsibility/social-impact-and-educational-initiatives /she-will-connect.html. 5. For more information, see Council of Europe 1985. 25. See http://aginghorizons.com/2015/01/program-young 6. See https://rankingdigitalrights.org. -immigrants-team-up-with-older-swedes/. 7. For information, see http://www.redipd.es. 26. See https://www.compisternli.ch/. 8. Children’s Online Privacy Protection Act of 1998, 15 USC 27. See https://e-estonia.com/solutions/education/e-school/. 6501–5. 28. An organization whose main activity is producing data- 9. Health Insurance Portability and Accountability Act of 1996 related products, services, and technologies. (HIPAA), Pub.L. 104–191, 110 Stat. 1936, enacted August 21, 1996. 29. See https://datazetu.or.tz/. Policies for the Data Economy 103 References Good Things Foundation. 2017. “Project: Reboot UK.” https:// www.goodthingsfoundation.org/projects/reboot-uk. Agrawal, Ravi. 2017. “India Supreme Court Rules Privacy a ‘Funda- Government of New Zealand. 2011. “New Zealand Data and mental Right’ in Landmark Case.” CNN.com, August 24. Information Management Principles.” Wellington. https:// https://www.cnn.com/2017/08/24/asia/indian-court-right-to www.ict .govt .nz/guidance-and-resources/open-government -privacy/index.html. /new-zealand-data-and-information-management-principles/. APEC (Asia-Pacific Economic Cooperation). 2015. The APEC Hesseldahl, Arik. 2015. “Why the Federal Government Sucks Privacy Framework. Singapore: APEC. at Cyber Security.” Recode, June 23. https://www.recode BBC (British Broadcasting Corporation). 2018. “Eric Schnei- .net/2015/6/23/11563798/why-the-federal-govern ment derman, New York Attorney General, Quits Amid Assault -sucks-at-cybersecurity. Reports.” BBC.com, May 8. https://www.bbc.com/news ITIF (Information Technology and Innovation Founda- /world-us-canada-44035718. tion). 2017. “Cross-Border Data Flows: Where Are the Burns, Thaddeus. 2017. “Regulating Machine Data: Less Is More Barriers, and What Do They Cost?” May 1. https://itif.org for Global Growth.” WIPO: Magazine, December. http://www / publications/2017/05/01/cross-border-data-flows-where .wipo.int/wipo_magazine/en/2017/06/article_0005.html. -are-barriers-and-what-do-they-cost. Chander, Anupam, and Uyen Lê. 2015. “Data Nationalism.” Kaminski, Piotr, Chris Rezek, Wolf Richter, and Marc Sorel. 2017. Emory Law Journal 64 (3): 677–739. http://law.emory.edu/elj “Protecting Your Critical Digital Assets: Not All Systems and /content/volume-64/issue-3/articles/data-nationalism.html. Data Are Created Equal.” McKinsey and Company. https:// Chang, Lulu. 2016. “The Latest Data Breach Involves the Voting www.mckinsey.com/business-functions/risk/our-insights Records of 93.4 Million Mexican Citizens.” Digital Trends, /protecting-your-critical-digital-assets-not-all-systems-and April 23. https://www.digitaltrends.com/computing/mexico -data-are-created-equal. -voting-breach/. Lindstedt, Catharina, and Daniel Naurin. 2005. “Transparency and Cisco. 2015. “Mitigating the Cybersecurity Skills Shortage: Top Corruption: The Conditional Significance of a Free Press.” Insights and Actions from Cisco Security Advisory Services.” Quality of Government Institute Working Paper 2005:5, https://www.cisco.com/c/dam/en/us/products/collateral/secur University of Gothenburg, Sweden. https://www.qog.pol.gu ity/cybersecurity-talent.pdf. .se/digitalAssets/1350/1350633_2005_5-lindstedt_naurin.pdf. Council of Europe. 1985. “Convention for the Protection of McKinsey Global Institute. 2016. The Age of Analytics: Competing Individuals with Regard to Automatic Processing of Personal in a Data-Driven World. London: McKinsey Global Institute. Data.” Council of Europe, Strasbourg. https://www.coe.int/en https://www.mckinsey.com/business-functions /mckinsey /web/conventions/full-list/-/conventions/treaty/108. -analytics/our-insights/the-age-of-analytics-competing-in-a Dhalla, Amira. 2016. “New Partnership with UN Women to Teach -data-driven-world. Key Digital Skills to Women.” Mozilla Learning. https:// Noack, Rick. 2018. “European Union’s New Privacy Law Made learning.mozilla.org/blog/new-partnership-with-un-women Some Websites Go Dark Today; Here’s What Else Has -to-teach-key-digital-skills-to-women. Changed.” Los Angeles Times, May 25. http://www.latimes European Commission. 2018. “New Measures to Boost Key .com/business/technology/la-fi-tn-gdpr-europe-privacy Competences and Digital Skills, as Well as the European -20180525-story.html. Dimension of Education.” Press release, January 17. http:// OECD (Organisation for Economic Co-operation and Develop- europa.eu/rapid/press-release_IP-18-102_en.htm. ment). 1980. “Rev. 2013 Guidelines on the Protection of European Union. 2012. “Charter of Fundamental Rights of the Privacy and Transborder Flows of Personal Data.” Paris. European Union.” 2012/C 326/02. http://eur-lex.europa.eu http://www.oecd.org/sti/ieconomy/privacy.htm. /legal-content/EN/TXT/?uri=CELEX:12012P/TXT. ———. 2015. Digital Security Risk Management for Economic ———. 2018. “Communication from the Commission to the and Social Prosperity: OECD Recommendation and Compan- European Parliament, the Council, the European Economic ion Document. Paris: OECD Publishing. https://doi and Social Committee and the Committee of the Regions .org/10.1787/9789264245471-en. ‘Towards A Common European Data Space.’” European Ohlhausen, Maureen K. 2017. “Remarks at the FTC Informa- Commission, Brussels. tional Injury Workshop.” December 12. https://www.ftc Facebook. 2017. “Global Government Request Report.” April 27. .gov /public-statements/2017/12/remarks-ftc-informational https://transparency.facebook.com/. -injury-workshop. Fell, Gregory, and Mike Barlow. 2016. “Not All Data Is Created Pfeifle, Sam. 2017. “Is the GDPR a Data Localization Law?” Inter- Equal.” O’Reilly, April 25. https://www.oreilly.com/learning national Association of Privacy Professionals, Portsmouth, NH. /not-all-data-is-created-equal. https://iapp.org/news/a/is-the-gdpr-a-data-localization-law/. 104 Information and Communications for Development 2018 Qatar Ministry of Information and Communications Technology. Virtu. 2015. “5 Data Security Challenges Faced by Government 2015. “Data Management Policy.” http://www.motc.gov.qa Agencies, and What They Can Do about It.” September 23. /sites/default/files/data_management_policy.pdf. https://www.virtru.com/blog/data-security/. Reuters. 2017. “Rwanda Regulator Fines MTN Rwanda $8.5 Mln Warren, Samuel D., and Louis D. Brandeis. 1890. “The Right to over External IT Hub.” May 17. https://www.reuters.com Privacy.” Harvard Law Review 4 (5): 193. /article/rwanda-telecoms/rwanda-regulator-fines-mtn-rwand World Bank. 2008. “P106589: Mexico Information Technology a-8-5-mln-over-external-it-hub-idUSL8N1IJ2IJ. (IT) Industry Development Project.” Washington, DC. http:// UNCTAD (United Nations Conference on Trade and Develop- projects.worldbank.org/P106589/information-technology ment). 2016. “Data Protection Regulations and International -development?lang=en. Data Flows: Implications for Trade and Development.” ———. 2014. “Open Data for Economic Growth.” Washington, UNCTAD, Geneva and New York. http://unctad.org/en DC. http://www.worldbank.org/content/dam /Worldbank /PublicationsLibrary/dtlstict2016d1_en.pdf. /document/Open-Data-for-Economic-Growth.pdf. ———. 2018. “Data Protection and Privacy Legislation Worldwide.” World Bank and CGAP (Consultative Group to Assist the Poor). Geneva and New York. http://unctad.org/en/Pages/DTL/STI_and 2018. “Data Protection and Privacy for Alternative Data.” Draft _ICTs/ICT4D-Legislation/eCom-Data-Protection-Laws.aspx. GPFI-FCPL Sub-Group Discussion Paper. Group of Twenty UNESCO (United Nations Educational, Scientific, and Cultural (G20). https://www.g20.org/sites/default/files/documentos Organization). 2018. A Lifeline to Learning: Leveraging Tech- _producidos/data_protection_and_privacy_for_alternative nology to Support Education for Refugees. Paris: UNESCO. _data_wbg_0.pdf. http://unesdoc.unesco.org/images/0026/002612/261278e.pdf. Wright, Morgan. 2018. “A Ransomware Attack Brought Atlanta USA Today. 2017. “Rebel Wilson Wins Defamation Case against to Its Knees—And No One Seems to Care.” The Hill, Publisher.” June 15. https://www.usatoday.com/story/life April 4. http://thehill.com/opinion/cybersecurity/381594-a /people/2017/06/15/rebel-wilson-wins-defamation-case -ransomware-attack-brought-atlanta-to-its-knees-and-no -against-publisher/102877120/. -one-seems-to. Policies for the Data Economy 105 Data Notes Data for Development Indicators T his appendix presents a series of statistics related Affordability and usage to different data categories. They measure data Price of 1 gigabyte (GB) of data refers to the lowest price for accessibility, affordability, usage, protection, infra- at least 1 GB per month of mobile data usage (table DN.1). structure, and availability of public sector data. The data Data are sourced from the Association for Affordable Inter- generally refers to 2016 or the latest available data. The net,4 the Organisation for Economic Co-operation and coverage and nomenclature shown for economy names Development (OECD),5 Research ICT Africa (RIA),6 the comes from the World Bank.1 Internet Society,7 and ictDATA.8 Data are shown in U.S. dollars and as a proportion of Availability and users gross domestic product (GDP) per capita. Data gener- These indicators refer to the technical availability of the ally refer to the largest operator by subscriber market potential to use data (mobile broadband coverage) and share. Note that the OECD data also includes 300 actual use (internet use). The proportion of the population voice calls and is generally postpaid, whereas data from covered by a mobile network is the sole indicator related to other sources is data only and generally prepaid. Note information and communication technology that is tracked that many operators provide more than 1 GB for the for the Sustainable Development Goals. The indicator used prices shown. Data usage is shown for two metrics: GB is the percentage of the population living within the signal per mobile data user and per all mobile subscriptions range of a third-generation (3G) mobile network and is regardless of whether data is used. The figures are for sourced from the United Nations.2 For actual usage, the monthly consumption. Data come from various sources, proportion of the population using the internet is shown, including the OECD,9 national agencies, and mobile sourced from the World Bank DataBank.3 operator groups. 107 Table DN.1 Data and affordability Availability and users Affordability and usage Proportion of Price of Price of population Individuals using 1 gigabyte 1 GB of data GB per covered by a 3G the internet (GB) of data (% of GDP GB per mobile mobile network, (% of population), (US dollars per per capita per data user, subscription, Economy 2015 2016 month), 2016 month), 2016 2016 2016 Afghanistan 40 10.6 4.41 9.4 0.323 0.75 Albania 99 66.4 4.20 1.2 Algeria 46 43.0 9.72 3.0 0.447 Angola 100 13.0 19.99 7.7 Antigua and Barbuda 98 73.0 25.92 2.2 Argentina 90 71.0 15.78 1.5 0.51 Armenia 100 67.0 4.14 1.4 Australia 99 88.2 22.82 0.5 1.543 Austria 98 84.3 13.35 0.4 6.278 Azerbaijan 97 78.2 2.91 0.9 Bahamas, The 98 80.0 20.00 1.0 Bahrain 98 98.0 15.79 0.8 Bangladesh 71 18.3 2.94 2.6 0.322 0.14 Barbados 98 79.6 13.50 1.0 Belarus 96 71.1 2.27 0.5 Belgium 100 86.5 18.33 0.5 0.863 Belize 44.6 15.00 3.7 Benin 45 12.0 11.28 17.2 0.339 0.07 Bhutan 80 41.8 3.06 1.3 Bolivia 27 39.7 7.42 2.9 Bosnia and Herzegovina 96 54.7 6.72 1.7 Botswana 92 39.4 28.15 5.0 Brazil 94 60.9 10.35 1.4 0.63 0.43 Brunei Darussalam 91 90.0 14.49 0.6 Bulgaria 100 59.8 5.74 0.9 Burkina Faso 10 14.0 8.05 14.9 Burundi 0 5.2 5.45 22.9 Cabo Verde 87 50.3 4.80 1.9 Cambodia 70 32.4 2.00 1.9 3.00 Cameroon 50 25.0 6.43 7.5 0.249 0.08 Canada 99 89.8 41.35 1.2 1.57 1.225 Central African Republic 23 4.0 6.58 20.7 Chad 13 5.0 16.45 29.7 Chile 90 66.0 28.14 2.5 1.2 China 95 53.2 19.92 2.9 Colombia 100 58.1 10.66 2.2 0.12 Comoros 7.9 8.68 13.4 Congo, Dem. Rep. 20 6.2 13.00 34.7 (continued next page) 108 Information and Communications for Development 2018 Table DN.1 (continued) Availability and users Affordability and usage Proportion of Price of Price of population Individuals using 1 gigabyte 1 GB of data GB per covered by a 3G the internet (GB) of data (% of GDP GB per mobile mobile network, (% of population), (US dollars per per capita per data user, subscription, Economy 2015 2016 month), 2016 month), 2016 2016 2016 Congo, Rep. 50 8.1 16.45 12.9 0.192 0.06 Costa Rica 99 66.0 16.71 1.7 Côte d’Ivoire 46 26.5 8.05 6.3 1.065 0.21 Croatia 99 72.7 7.55 0.7 Cuba 0 38.8 Cyprus 90 75.9 11.30 0.6 2.827 1.47 Czech Republic 99 76.5 31.74 2.1 0.983 Denmark 100 97.0 23.14 0.5 4.373 Djibouti 0 13.1 45.01 Dominica 60 67.0 17.41 2.9 Dominican Republic 99 61.3 8.68 1.6 Ecuador 92 54.1 20.00 4.0 Egypt, Arab Rep. 98 41.3 1.36 0.5 0.31 El Salvador 73 29.0 5.00 1.4 Equatorial Guinea 23.8 Eritrea 92 1.2 Estonia 100 87.2 10.88 0.7 4.127 Ethiopia 71 15.4 7.44 12.6 Fiji 68 46.5 11.92 2.8 Finland 100 87.7 22.80 0.6 10.948 France 99 85.6 17.20 0.6 1.618 Gabon 97 48.1 8.79 1.5 Gambia, The 86 18.5 6.55 16.6 Georgia 99 58.0 1.99 0.6 Germany 96 89.7 28.95 0.8 1.212 Ghana 80 34.7 4.73 3.8 0.282 0.15 Greece 99 69.1 59.69 4.0 0.718 Grenada 75 55.9 16.67 2.1 Guatemala 92 34.5 13.26 3.8 Guinea 39 9.8 3.30 7.8 0.157 0.07 Guinea-Bissau 3.8 58.29 112.8 0.054 0.02 Guyana 0 35.7 9.69 2.6 Haiti 50 12.2 3.84 6.2 Honduras 83 30.0 17.47 8.9 Hong Kong SAR, China 99 87.5 6.16 0.2 1.477 1.602 Hungary 99 79.3 44.77 4.2 1.423 Iceland 99 98.2 24.47 0.5 3.921 India 74 29.6 3.77 2.7 0.88 (continued next page) Data for Development Indicators 109 Table DN.1 (continued) Availability and users Affordability and usage Proportion of Price of Price of population Individuals using 1 gigabyte 1 GB of data GB per covered by a 3G the internet (GB) of data (% of GDP GB per mobile mobile network, (% of population), (US dollars per per capita per data user, subscription, Economy 2015 2016 month), 2016 month), 2016 2016 2016 Indonesia 60 25.4 4.10 1.4 Iran, Islamic Rep. 60 53.2 3.91 0.9 1.112 0.6 Iraq 55 21.2 12.67 3.3 Ireland 95 85.0 34.36 0.7 3.1 Israel 99 79.7 21.85 0.7 Italy 100 61.3 28.70 1.1 1.672 Jamaica 90 45.0 15.31 3.8 Japan 100 93.2 70.72 2.2 2.121 Jordan 99 62.3 7.08 2.1 Kazakhstan 73 74.6 4.48 0.7 Kenya 69 26.0 4.98 4.1 Kiribati 63 13.7 37.56 31.1 Korea, Rep. 99 92.8 17.69 0.8 3.833 Kosovo 91 82.9 5.65 1.8 Kuwait 97 78.4 16.67 0.7 Kyrgyz Republic 59 34.5 1.74 1.9 Lao PDR 65 21.9 5.99 3.1 Latvia 95 79.8 10.30 0.9 8.21 Lebanon 97 76.1 19.00 2.9 Lesotho 96 27.4 7.19 8.6 Liberia 50 7.3 5.00 13.2 0.178 0.06 Libya 50 20.3 10.80 Lithuania 100 74.4 4.52 0.4 2.51 1.37 Luxembourg 99 98.1 22.91 0.3 2.912 Macao SAR, China 100 81.6 12.00 0.2 Macedonia, FYR 98 72.2 3.64 0.8 Madagascar 61 4.7 4.67 14.0 Malawi 32 9.6 5.36 21.4 Malaysia 92 78.8 6.73 0.8 3.1 1.92 Maldives 100 59.1 12.95 1.8 Mali 10 11.1 12.08 18.6 0.179 0.038 Malta 100 77.3 13.56 0.6 Marshall Islands 29.8 Mauritania 30 18.0 11.71 13.0 Mauritius 93 52.2 8.62 1.1 Mexico 89 59.5 11.59 1.7 0.27 Micronesia, Fed. Sts. 33.4 30.00 11.7 Moldova 99 71.0 2.70 1.7 (continued next page) 110 Information and Communications for Development 2018 Table DN.1 (continued) Availability and users Affordability and usage Proportion of Price of Price of population Individuals using 1 gigabyte 1 GB of data GB per covered by a 3G the internet (GB) of data (% of GDP GB per mobile mobile network, (% of population), (US dollars per per capita per data user, subscription, Economy 2015 2016 month), 2016 month), 2016 2016 2016 Mongolia 95 22.3 6.76 2.2 Montenegro 97 69.9 22.93 4.1 Morocco 80 58.3 4.99 2.1 Mozambique 50 17.5 2.74 8.6 Myanmar 79 25.1 2.14 2.0 Namibia 37 31.0 13.42 3.9 Nauru 98 22.30 3.4 Nepal 50 19.7 2.80 4.6 0.177 0.07 Netherlands 99 90.4 30.58 0.8 1.024 New Zealand 98 88.5 28.49 0.9 1.057 Nicaragua 75 24.6 13.87 7.7 Niger 10 4.3 6.58 21.7 Nigeria 67 25.7 3.21 1.8 0.164 0.08 Norway 99 97.3 25.12 0.4 2.594 Oman 95 69.9 13.16 1.1 Pakistan 46 15.5 1.54 1.3 0.464 Palau 88 49.90 4.4 Panama 79 54.0 15.00 1.3 Papua New Guinea 60 9.6 21.70 10.4 Paraguay 66 51.4 13.35 3.9 Peru 91 45.5 17.87 3.6 0.42 Philippines 78 55.5 6.02 2.4 Poland 100 73.3 10.52 1.0 3.548 Portugal 99 70.4 34.93 2.1 1.521 Qatar 98 94.3 16.48 0.3 Romania 100 59.5 4.48 0.6 0.81 Russian Federation 73 73.1 6.86 0.9 2.315 1.43 Rwanda 88 20.0 4.91 8.4 0.271 0.12 Samoa 86 29.4 9.36 2.8 São Tomé and Príncipe 2 28.0 8.84 6.0 Saudi Arabia 97 73.8 29.33 1.8 Senegal 40 25.7 3.06 3.8 0.709 0.198 Serbia 97 67.1 3.67 0.8 Seychelles 90 56.5 18.33 1.5 Sierra Leone 20 11.8 17.87 43.2 Singapore 100 81.0 7.24 0.2 Slovak Republic 93 80.5 32.41 2.4 0.66 Slovenia 100 75.5 19.84 1.1 1.419 (continued next page) Data for Development Indicators 111 Table DN.1 (continued) Availability and users Affordability and usage Proportion of Price of Price of population Individuals using 1 gigabyte 1 GB of data GB per covered by a 3G the internet (GB) of data (% of GDP GB per mobile mobile network, (% of population), (US dollars per per capita per data user, subscription, Economy 2015 2016 month), 2016 month), 2016 2016 2016 Solomon Islands 12 11.0 25.16 15.1 Somalia 30 1.9 South Africa 98 54.0 10.75 2.5 0.44 0.26 South Sudan 20 6.7 133.70 0.084 0.02 Spain 99 80.6 38.75 1.8 Sri Lanka 83 32.1 1.36 0.4 2.1 0.55 St. Kitts and Nevis 100 76.8 17.03 1.2 St. Lucia 65 46.7 14.81 2.3 St. Vincent and the Grenadines 100 55.6 14.81 2.5 Sudan 46 28.0 4.64 2.3 0.361 0.19 Suriname 100 45.4 19.77 3.7 Swaziland 21 28.6 32.70 14.1 0.16 0.09 Sweden 100 89.7 18.05 0.4 4.383 Switzerland 100 89.1 31.40 0.5 2.712 Syrian Arab Republic 70 31.9 8.12 0.165 0.06 Taiwan, China 100 86.3 6.55 0.3 7.4 Tajikistan 60 20.5 1.52 2.3 Tanzania 85 13.0 4.68 6.4 Thailand 97 47.5 5.57 1.1 3.9 Timor-Leste 96 25.3 15.00 12.8 Togo 50 11.3 8.23 17.1 Tonga 70 40.0 10.83 3.5 Trinidad and Tobago 75 73.3 22.34 1.7 Tunisia 94 49.6 4.35 1.4 Turkey 95 58.4 6.65 0.7 2.22 Turkmenistan 60 18.0 14.29 2.7 Tuvalu 46.0 Uganda 60 21.9 8.43 16.4 0.144 0.05 Ukraine 35 52.5 2.82 1.5 United Arab Emirates 100 90.6 27.25 0.9 United Kingdom 100 94.8 21.91 0.7 1.839 United States 100 76.2 46.62 1.0 2.665 Uruguay 90 66.4 5.23 0.4 0.79 Uzbekistan 32 46.8 4.94 2.8 Vanuatu 51 24.0 9.22 3.9 0.379 0.104 Venezuela, RB 90 60.0 1.05 0.35 Vietnam 70 46.5 1.78 1.0 (continued next page) 112 Information and Communications for Development 2018 Table DN.1 (continued) Availability and users Affordability and usage Proportion of Price of Price of population Individuals using 1 gigabyte 1 GB of data GB per covered by a 3G the internet (GB) of data (% of GDP GB per mobile mobile network, (% of population), (US dollars per per capita per data user, subscription, Economy 2015 2016 month), 2016 month), 2016 2016 2016 West Bank and Gaza 0 61.2 Yemen, Rep. 80 24.6 11.61 14.1 0.103 0.02 Zambia 53 25.5 13.27 13.5 0.217 0.08 Zimbabwe 55 23.1 35.00 41.6 East Asia and Pacific 82 51.6 16.34 4.2 2.8 1.7 Europe and Central Asia 92 75.2 16.18 1.3 2.9 1.3 Latin America and the 81 56.2 14.30 2.8 0.6 0.5 Caribbean Middle East and North Africa 77 58.0 14.64 2.2 0.5 0.2 North America 99 88.0 43.99 1.1 2.1 1.2 South Asia 68 28.3 4.10 3.0 0.7 0.5 Sub-Saharan Africa 52 20.1 14.05 14.9 0.3 0.1 Low income 43 11.2 14.59 20.7 0.24 0.13 Lower middle income 65 33.6 9.46 5.4 0.44 0.38 Upper middle income 82 53.0 11.50 2.2 1.77 0.65 High income 96 82.5 23.63 1.1 2.91 1.28 World 77 51.7 14.98 5.5 1.82 0.51 Note: Data for groups is compiled using averages. Blank cells indicate that no information is available for the indicator. Government These indicators refer to aspects of government involvement instances, no specific law exists on data privacy and protec- with data. Data are provided on the number of open data tion and instead the principles are presumably encapsulated sets in the economy and the existence of a data protection in other laws, such as a constitution or electronic trans- and privacy law, as well as for the authority responsible for action act. It is to be noted that, in some instances, the listed data protection (table DN.2). The number of open data sets statutes refer to draft legislation. is sourced from OpenDataSoft (Mercier 2015). No definitive list of data protection authorities exists. Data have been Infrastructure sourced from the International Conference of Data Protec- tion and Privacy Commissioners.10 The data should there- These indicators refer to the potential for domestic data fore be treated with caution, as there may be cases in which exchange and the volume of international internet band- a data protection authority exists in an economy but is not a width to provide an indication of the quality of data member of the International Conference of Data Protection transmission. Data are provided on the number of internet and Privacy Commissioners. Data protection and privacy exchange points and international internet bandwidth. The laws have been sourced from the United Nations Conference number of internet exchange points is sourced from Packet on Trade and Development.11 In cases in which economies Clearing House.12 Data on international internet bandwidth have reported multiple laws, the one specifically referring is sourced from the International Telecommunications to data privacy or protection is listed. Note that, in some Union (ITU 2017). Data for Development Indicators 113 Table DN.2 Government data infrastructure and open data Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Afghanistan 3 0.09 0 11,967 Albania Commissioner for Law No. 9887 on the 3 1.04 1 56,964 Personal Data Protection Protection of Personal Data Algeria 5 0.12 0 40,015 American Samoa Andorra Data Protection Agency Loi qualifiée 15/2003, 2 25.99 0 106,390 du 18 décembre, sur la protection des données personnelles Angola Lei No. 22/11 da 2 0.07 2 8,796 Protecção de Dados Pessoais de 17 de Junho (in Portuguese) Antigua and Barbuda Data Protection Act 2013 1 9.90 0 88,622 Argentina National Direction for Ley 25.326 de 23 0.52 27 41,130 Personal Data Protection Protección de los Datos Personales Armenia Personal Data Protection Law of the Republic of 3 1.03 1 59,860 Agency Armenia on the Protection of Personal Data Aruba 0 0.00 0 Australia Office of the Privacy Act 1988 68 2.82 19 88,304 Australian Information Commissioner Austria Austrian Data Protection Datenschutzgesetz 2000 20 2.29 4 149,988 Authority Azerbaijan Law on Personal Data 3 0.31 0 34,255 2010 Bahamas, The Data Protection (Privacy 1 2.56 0 198,447 of Personal Information) Act 2003 Bahrain 3 2.11 1 112,770 Bangladesh 4 0.02 1 9,154 Barbados Data Protection Bill 2005 1 3.51 0 284,571 Belarus Law of the Republic of 4 0.42 0 168,518 Belarus on Information, Informatization and Protection of Information— Law no. 8517 Belgium Privacy Commission Law on Privacy 28 2.47 1 189,254 Protection in relation to the Processing of Personal Data Belize 3 8.18 1 44,633 (continued next page) 114 Information and Communications for Development 2018 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Benin National Commission Loi n°2009-09 du 22 4 0.37 1 1,656 for Technology and mai 2009 portant Freedoms organisation de la protection des données à caractère personnel Bermuda 3 45.92 Bhutan Bhutan Information 0 0.00 0 18,077 Communications and Media Act 2006 Bolivia Ley General de 30 2.76 2 36,347 Telecomunicaciones, Tecnologías de Información y Comunicación—Ley 167 de 08 agosto de 2011 Bosnia and Personal Data Protection Law on the Protection of 4 1.14 0 98,452 Herzegovina Agency in Bosnia and Personal Data Herzegovina Botswana 2 0.89 1 7,880 Brazil Protection of Personal 23 0.11 27 66,181 Data Bill 2011 British Virgin Islands 0 0 Brunei Darussalam 2 4.73 0 76,226 Bulgaria Commission for Personal Law for Protection of 6 0.84 5 175,869 Data Protection Personal Data Burkina Faso Data Processing and Loi n° 010-2004/AN 6 0.32 1 2,810 Liberties Commission Portant Protection des Données à Caractère Personnel Burundi 3 0.29 1 6,083 Cabo Verde National Commission of Lei n° 133/V/2001 of 22 4 7.41 0 23,357 Data Protection January 2001 Cambodia 2 0.13 2 23,573 Cameroon 7 0.30 0 2,549 Canada Privacy Commissioner of Personal Information 159 4.38 13 141,885 Canada (Commissariat Protection and Electronic à la protection de la vie Documents Act privée du Canada) Cayman Islands 1 16.46 Central African 4 0.87 0 1,695 Republic Chad Law 007/PR/2015 on the 1 0.07 0 3,762 Protection of Personal Data Channel Islands 0 0.00 (continued next page) Data for Development Indicators 115 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Chile Law 19.628 of 1999 14 0.78 6 175,556 China The Decision of the 16 0.01 8 14,699 Standing Committee of the National People’s Congress on Strengthening the Network Information Protection (2012) Colombia Superintendence of Law 1266 of 2008- 38 0.78 1 150,871 Industry and Commerce Habeas Data Act of Colombia Comoros 1 1.26 0 12,729 Congo, Dem. Rep. 5 0.08 1 770 Congo, Rep. 6 1.17 1 Costa Rica Agency for the Ley de Protección de 13 2.68 1 68,449 Protection of Personal la Persona frente al Data of Inhabitants tratamiento de sus datos personales, Nº 8968 Côte d’Ivoire Telecommunications/ICT Loi n° 2013-450 du 5 0.21 1 6,825 Regulatory Body (ARTCI) 19 juin 2013 relative à la protection des données à caractère personnel Croatia Data Protection Agency Act on Personal Data 9 2.16 1 118,953 Protection Cuba 3 0.26 1 1,152 Curaçao 0 0 1 Cyprus Personal Data Protection The Processing of 4 3.42 1 188,904 Commissioner Personal Data (Protection of the Individual) Law Czech Republic Office for Personal Data Law on Personal Data 21 1.99 3 180,697 Protection Protection Denmark Data Protection Agency Act on Processing of 18 3.14 3 239,874 Personal Data Djibouti 2 2.12 1 15,228 Dominica Privacy and Data 1 13.60 1 176,449 Protection Bill 2007 Dominican Republic Ley No. 172-13, sobre 4 0.38 1 22,061 Protección de Datos de Carácter Personal del 13 de diciembre de 2013 Ecuador Protection of Privacy and 4 0.24 4 43,677 Personal Data Bill 2016 Egypt, Arab Rep. 8 0.08 2 17,194 El Salvador Ley de Comercio 6 0.95 0 63,622 Electronico y Comunicaciones (continued next page) 116 Information and Communications for Development 2018 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Equatorial Guinea Law 1/2016 (Data 3 2.46 0 2,397 protection law) Eritrea 1 0.19 0 3,601 Estonia Data Protection Data Protection Act 5 3.80 3 210,798 Inspectorate Ethiopia 4 0.04 0 2,242 Faroe Islands 0 0.00 Fiji 2 2.23 0 23,726 Finland Data Protection Personal Data Act 26 4.73 4 216,391 Ombudsman France National Commission Law relating to the 155 2.32 16 97,653 of Computing and protection of individuals Freedoms (CNIL) against the processing of personal data French Polynesia 0 0.00 Gabon Loi n°001/2011 relative 5 2.53 1 4,844 à la protection des données à caractère personnel Gambia, The Information and 2 0.98 1 13,297 Communications Act No. 2 of 2009 Georgia Office of the Personal Law of Georgia on 3 0.81 1 92,145 Data Protection Personal Data Protection Inspector of Georgia Germany Federal Data Protection Federal Data Protection 57 0.69 19 107,489 Commissioner Act Ghana Data Protection Data Protection Act 4 0.14 1 9,851 Commission (GDPC) (Act No. 843) 2012—DPA Gibraltar Data Protection 0 0 Commissioner Greece Hellenic Data Protection Law on the Protection of 5 0.47 1 68,698 Authority individuals with regard to the processing of personal data Greenland 0 0.00 0 Grenada 0 0.00 1 229,948 Guam 0 0.00 Guatemala 4 0.24 0 24,022 Guinea 3 0.24 0 589 Guinea-Bissau 2 1.10 0 4,707 Guyana 3 3.88 0 34,675 Haiti 3 0.28 1 2,337 (continued next page) Data for Development Indicators 117 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Honduras Anteproyecto de Ley 1 0.11 1 33,443 de Protección de Datos Personales y Acción de Hábeas Data de Honduras Hong Kong SAR, Privacy Commissioner 0 0.00 4,906,023 China for Personal Data Hungary National Authority for Act on Informational 4 0.41 1 154,765 Data Protection and Self-Determination and Freedom of Information Freedom of Information Iceland Data Protection Law on the Protection 8 23.93 1 997,830 Authority and Processing of Personal Data 1989 India Information Technology 6 0.00 13 15,956 Act 2000 Indonesia Law of the Republic 8 0.03 7 24,947 of Indonesia Number 11 of 2008 Concerning Electronic Information and Transactions Iran, Islamic Rep. Law on Electronic 3 0.04 4 15,238 Commerce Iraq Draft Data Protection 0 0.00 0 and Privacy Law Ireland Data Protection Data Protection Act, 15 3.14 3 183,943 Commissioner 1988 Isle of Man Data Protection 0 0 Registrar Israel The Israeli Law, Privacy Protection Act 8 0.94 1 158,696 Information and 1981 Technology Authority Italy Data Protection Decreto Legislativo 56 0.92 8 82,335 Commission 30 giugno 2003, n. 196— Codice in materia di protezione dei dati personali Jamaica Data Protection Bill 2012 5 1.74 1 47,949 Japan Personal Information Act on the Protection of 20 0.16 16 83,010 Protection Commission Personal Information Jordan Data Protection Bill 6 0.63 0 8,229 Kazakhstan Law on personal data 0 0.00 0 87,235 and their protection, 21 May 2013 Kenya Data Protection Bill 2012 9 0.19 2 69,014 Kiribati 0 0.00 0 4,426 Korea, Dem. People’s 0 0.00 0 Rep. (continued next page) 118 Information and Communications for Development 2018 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Korea, Rep. Personal Information Korea’s Personal 10 0.20 3 54,252 Protection Commission Information Protection Act was promulgated in 2011 as amended Kosovo National Agency 2 1.10 1 for Personal Data Protection, AMDP Kuwait Law No. 20 of 2014 8 1.97 0 69,516 Kyrgyz Republic Law on Personal Data 2 0.33 1 65,377 2008 Lao PDR 3 0.44 1 17,487 Latvia State Data Inspectorate Law on Protection of 3 1.53 2 246,666 Personal Data of Natural Persons Lebanon 1 0.17 2 55,086 Lesotho Data Protection Act 2012 4 1.82 1 4,484 Liberia 1 0.22 1 Libya 2 0.32 0 5,286 Liechtenstein Data Protection Gesetz über die 0 0.00 1 Commissioner Abänderung des Datenschutzgesetzes, 2002 Lithuania State Data Inspectorate Law on Legal Protection 10 3.48 3 198,564 of Personal Data Luxembourg National Data Protection Data Protection Law 3 5.15 1 8,397,884 Commission Macao SAR, China 0 0.00 252,868 Macedonia, FYR Directorate of Personal Law on Personal Data 3 1.44 0 109,004 Data Protection Protection Madagascar Loi No. 2014-38 3 0.12 1 14,258 Malawi Electronic Transactions and 5 0.28 1 4,201 Cybersecurity Act 2016 Malaysia Personal Data Protection 4 0.13 1 42,627 Act 2010 Maldives 2 4.79 0 59,669 Mali Personal Data Protection Loi n° 2013-015 du 21 5 0.28 1 598 Authority (Autorité de mai 2013 Protection de Données à Caractère Personnel) Malta Data Protection Data Protection Act 2001 3 6.87 1 1,596,254 Commissioner Marshall Islands 0 0.00 0 Mauritania 2 0.47 0 4,477 Mauritius Data Protection Office of Data Protection Act 3 2.37 1 63,491 Mauritius 2004 (continued next page) Data for Development Indicators 119 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Mexico National Institute for Ley Federal de 23 0.18 1 37,598 Transparency, Access to Protección de Datos Information and Personal Personales en Posesión Data Protection de los Particulares 2010 Micronesia, Fed. Sts. 2 19.06 0 Moldova National Center for Law on Personal Data 5 1.41 0 144,087 Personal Data Protection Protection 2007 Monaco Supervisory Commission Act controlling personal 1 25.97 0 95,232 for Personal Information data processing 1993 Mongolia 2 0.66 1 166,056 Montenegro Agency for Personal Law on Personal Data 3 4.82 1 202,876 Data Protection and Free Protection 2008 Access to Information Morocco National Commission Law No. 09-08/2009 on 4 0.11 0 25,702 for the Control and the the protection of people Protection of Personal toward data protection Data of a personal nature Mozambique 3 0.10 1 1,115 Myanmar 5 0.09 1 6,426 Namibia 4 1.61 1 15,915 Nauru 1 76.63 0 Nepal Right to Information Act, 8 0.28 1 3,886 2064 (2007) Netherlands Dutch Data Protection Personal Data Protection 53 3.11 8 196,105 Authority Act 1998 New Caledonia 3 10.79 0 New Zealand Privacy Privacy Act 1993 19 4.05 7 109,601 Commissioner (Te Mana Matapono Matatapu) Nicaragua Ley No. 787 Ley de 2 0.33 1 29,161 Protección de Datos Personales Niger Projet de loi sur la 3 0.15 0 protection des données à caractère personnel Nigeria Data Protection Bill 2011 4 0.02 2 11,257 Northern Mariana 0 0.00 Islands Norway Data Inspectorate Personal Data Act 2000 6 1.15 5 268,953 Oman Royal Decree no. 69 2 0.45 0 66,071 of 2008—Electronic Transactions Law Pakistan Electronic Data Protection 6 0.03 1 16,636 Act 2005—Draft (continued next page) 120 Information and Communications for Development 2018 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Palau 1 46.51 0 Panama Personal Data Protection 4 0.99 1 55,072 Bill 2016 Papua New Guinea 2 0.25 1 Paraguay Ley 1682/2001 4 0.59 1 21,015 Reglamenta la Informacion de Caracter Privado Peru National Authority for Ley N° 29733—Ley de 19 0.60 1 33,315 Data Protection Protección de Datos Personales Philippines National Privacy Data Privacy Act of 2012 5 0.05 1 43,440 Commission Poland Inspector General for Act on the Protection of 6 0.16 10 83,299 Personal Data Protection Personal Data 1997 Portugal National Data Protection Lei da proteçao de 6 0.58 1 177,808 Commission dados pessoais 1991 Puerto Rico 0 0.00 1 Qatar Law No. 13 of 2016 2 0.78 1 86,950 Concerning Privacy and Protection of Personal Data Romania National Supervisory Law on the protection of 6 0.30 3 155,516 Authority for Personal individuals with regard Data Protection to the processing of personal data etc. (2001) Russian Federation Federal Law Regarding 8 0.06 27 51,888 Personal Data 2006 Rwanda 4 0.34 1 7,455 Samoa 3 15.37 0 13,159 San Marino Law regulating the 1 30.12 0 Computerized Collection of Personal Data 1983 São Tomé and Data Protection Law 2 10.00 0 37,317 Príncipe 2016 Saudi Arabia 10 0.31 0 78,163 Senegal Commission of Personal Loi n° 2008-12 du 25 3 0.19 1 4,977 Data Protection, CDP janvier 2008 sur la protection des données à caractère personnel Serbia Commissioner for Law on Personal Data 4 0.57 1 26,292 Information of Public Protection 2008 Importance and Personal Data Protection Seychelles Data Protection Act 4 42.25 0 52,433 2003 (continued next page) Data for Development Indicators 121 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Sierra Leone 3 0.41 0 Singapore Personal Data Protection Personal Data Protection 4 0.71 5 982,923 Commission Act 2012 Sint Maarten 1 (Dutch part) Slovak Republic Inspection Unit for the Act on the Protection of 6 1.11 3 52,351 Protection of Personal Personal Data 1992 Data Slovenia Information Personal Data Protection 8 3.87 1 239,168 Commissioner of the Act 1990 Republic of Slovenia Solomon Islands 2 3.34 0 11,971 Somalia 3 0.21 0 South Africa Information Regulator Protection of Personal 12 0.21 6 263,030 Information Act 4 of 2013 South Sudan 2 0.16 0 Spain Data Protection Agency Organic Law 15/1999 on 72 1.55 5 112,997 Personal Data Protection Sri Lanka 2 0.09 1 22,038 St. Kitts and Nevis Privacy and Data 0 0.00 0 165,372 Protection Bill 2012 St. Lucia Data Protection Act 2011 1 5.62 1 7,558 St. Martin (French part) St. Vincent and the Privacy Act 2003 0 0.00 0 188,740 Grenadines Sudan 4 0.10 1 2,035 Suriname The Constitution of the 3 5.37 0 66,533 Republic of Suriname, Article 17 Swaziland Data Protection Bill 3 2.23 0 Sweden Data Inspection Board Personal Data Act 1998 33 3.33 9 505,650 Switzerland Federal Data Protection Federal Act on Data 16 1.91 3 269,222 Commissioner Protection, 1992 Syrian Arab Republic 0 0.00 0 12,813 Taiwan, China Personal Data Protection 11 0.47 3 97,652 Law Tajikistan Law on Protection of 2 0.23 0 Information (December 2, 2002, ʋ 71) Tanzania Data Protection Bill 2013 7 0.13 3 1,741 Thailand Personal Data Protection 16 0.23 2 49,244 Bill 2011 Timor-Leste 3 2.36 1 1,888 (continued next page) 122 Information and Communications for Development 2018 Table DN.2 (continued) Government Data infrastructure International Open data Number internet Number of sets per of internet bandwidth (bits/s Data protection Data privacy and open data million exchange per internet Economy authority protection legislation sets, 2018 people points, 2018 user), 2016 Togo 2 0.26 1 4,490 Tonga 2 18.67 0 33,947 Trinidad and Tobago Data Protection Act 2011 2 1.47 2 182,808 Tunisia National Personal Data Law 63/2004 10 0.88 2 31,167 Authority Turkey Personal Data Protection Data Protection Law 4 0.05 2 68,058 Authority 2016 Turkmenistan 1 0.18 0 Tuvalu 2 180.23 0 Uganda The Data Protection and 7 0.17 1 5,510 Privacy Bill, 2015 Ukraine Ukrainian Parliament Law on Personal Data 8 0.18 6 79,885 Commissioner for Protection 2011 Human Rights United Arab Emirates Dubai International 3 0.32 2 133,749 Financial Centre (DIFC) Data Protection Law United Kingdom Information Data Protection Act 154 2.35 9 449,137 Commissioner’s Office 1998 United States Federal Trade Privacy Act of 1974 993 3.07 89 126,545 Commission Uruguay Regulatory and Control La Ley 18331 Protección 7 2.03 0 96,707 Unit of Personal Data de Datos Personales y Acción de Habeas Data”del 11 agosto del año 2008 y el Decreto reglamentario 414/2009 Uzbekistan 2 0.06 1 5,683 Vanuatu 2 7.40 1 21,921 Venezuela, RB 3 0.10 0 18,937 Vietnam Law on Protection of 4 0.04 3 91,252 Consumers’ Rights 2010 Virgin Islands (U.S.) 0 West Bank and Gaza 5 1.10 1 0 Yemen, Rep. Law of the Right of Access 2 0.07 0 to Information 2012 Zambia The Electronic 2 0.12 1 3,925 Communications and Transactions Act, Act Number 21 of 2009—the Electronic Communications Act Zimbabwe Draft Data Protection 2 0.12 1 9,119 Bill 2016 Note: Blank cells indicate that no information is available for this country on this indicator. Data for Development Indicators 123 Digital Adoption Index The DAI is intended to reflect actual adoption of digital The Digital Adoption Index is a global index that measures technologies across the economy, not perceptions of adop- countries’ digital adoption across three dimensions of the tion. Accordingly, indicators comprising the index represent economy: people, government, and business. The index subscriptions, access, or adoption and eschew public or covers 180 countries on a 0–1 scale and emphasizes the elite opinion surveys. Most data come from the Inter- “supply side” of digital adoption to maximize coverage and national Telecommunication Union or the World Bank. simplify theoretical linkages (map DN.1). The overall index Other sources include Eurostat, GSMA, and Netcraft (table is the simple average of three subindexes. Each subindex DN.3). Data were collected for two time periods: 2014 and comprises technologies necessary for the respective agent to 2016. The business subindex measures the quality of digital promote development in the digital era: increasing productiv- infrastructure needed for e-commerce and other business ity and accelerating broad-based growth for business, expand- functions, comprising the number of secure servers and ing opportunities and improving welfare for people, and international internet bandwidth, as well as the percentage increasing the efficiency and accountability of service delivery of businesses with websites as a proxy for their more general for government. Although data and theoretical constraints online business activities. The people subindex measures the prohibit any index from providing a comprehensive view extent and quality of individuals’ connections to the digital of an economy, the Digital Adoption Index (DAI) provides world, comprising access to mobile-cellular phones, basic a useful framing mechanism for digital adoption across internet, and mobile and fixed broadband. And the govern- economic agents and countries. By measuring the relative ment subindex measures the adoption of core administrative adoption of digital technologies, the index can assist policy systems to automate and streamline government activities makers in designing a digital strategy with tailored policies to and digital identification systems and online public services promote digital adoption across different user groups.13 that allow the government to better serve the public. Map DN.1 Digital Adoption Index score, by country Digital adoption index score by country 0 – 0.14 0.15 – 0.29 0.30 – 0.43 0.44 – 0.58 0.58 – 0.72 0.73 – 0.87 No data IBRD 43805 | SEPTEMBER 2018 Source: World Bank 2018. Note: The Digital Adoption Index measures digital adoption for 180 countries. 124 Information and Communications for Development 2018 Table DN.3 Data sources for the Digital Adoption merely a reordering based on relative trajectories between Index countries.14 For example, the DAI score for India increased Subindex Indicator Source from 0.44 to 0.51 over the period 2014–16. This means Business Internet bandwidth ITU that India made progress over the period, increasing digital Business websites Eurostat and World Bank adoption. If scores were normalized within years (not across Secure servers Netcraft years), it would not be possible to determine if an increase in People Mobile-cellular ITU India’s score meant India made progress or other countries subscriptions regressed. Mobile broadband GSMA Normalized data is averaged so that the DAI’s constitu- Internet use ITU ent indicators have equal weight at each level. For example, Fixed broadband ITU the government indicator of core administrative systems Government Core administrative World Bank is composed of four categorical variables collected by the systems World Bank. After the scores are normalized, the simple Digital identification World Bank average is calculated. The resulting average represents the Online public services UNDESA country’s score for core administrative systems. The core Source: World Bank 2018. administrative systems score is then averaged with the scores Note: ITU = International Telecommunication Union; UNDESA = United Nations Department of Economic and Social Affairs. for digital identification and online public services, which are generated using a similar process. And the resulting average at that level represents the government subindex Missing values are estimated and indicators normalized to score. The business and people subindexes follow the same create a complete and balanced data set. A few indicators— process. The overall DAI varies on a 0–1 scale because its particularly the percentage of businesses with websites—are source indicators are normalized to that scale. As with the missing observations for many countries. Instead of drop- subindexes, 0 is the lowest possible score on the DAI, repre- ping the countries, missing values are imputed using data senting no adoption of digital technologies, and 1 is the on per capita income, internet use, and geographical region. highest possible score, representing full adoption of digital Data are normalized on a 0–1 scale so that each indicator has technologies. Theoretically, a country can score a perfect 1 if equal weight within a subindex. In all but one case, indica- it has the best score on all the indicators comprising the DAI tors are normalized across both years, not within each indi- or a perfect 0 if it has the worst score on all the indicators. vidual year. Observed changes can therefore be considered But in practice, DAI scores ranged 0.14–0.87 in 2014 and absolute changes in value for particular countries, not 0.15–0.87 in 2016 (table DN.4). Table DN.4 Digital Adoption Index and subindexes, by country and year Digital Adoption Index Business subindex People subindex Government subindex Country 2014 2016 2014 2016 2014 2016 2014 2016 Afghanistan 0.30 0.34 0.30 0.34 0.09 0.12 0.52 0.56 Albania 0.54 0.61 0.55 0.62 0.39 0.46 0.69 0.74 Algeria 0.37 0.43 0.45 0.50 0.28 0.42 0.38 0.38 Andorra 0.59 0.64 0.78 0.83 0.67 0.74 0.33 0.35 Angola 0.32 0.33 0.38 0.41 0.12 0.13 0.45 0.46 Antigua and Barbuda 0.46 0.48 0.60 0.61 0.45 0.57 0.33 0.25 Argentina 0.64 0.69 0.68 0.69 0.56 0.63 0.68 0.73 (continued next page) Data for Development Indicators 125 Table DN.4 (continued) Digital Adoption Index Business subindex People subindex Government subindex Country 2014 2016 2014 2016 2014 2016 2014 2016 Armenia 0.61 0.62 0.68 0.71 0.41 0.48 0.73 0.67 Australia 0.69 0.71 0.76 0.77 0.72 0.78 0.57 0.59 Austria 0.81 0.86 0.84 0.88 0.82 0.87 0.79 0.85 Azerbaijan 0.55 0.59 0.47 0.51 0.51 0.52 0.67 0.75 Bahamas, The 0.48 0.53 0.68 0.74 0.43 0.50 0.32 0.35 Bahrain 0.76 0.79 0.70 0.75 0.77 0.84 0.81 0.77 Bangladesh 0.31 0.37 0.30 0.36 0.15 0.19 0.48 0.57 Barbados 0.57 0.65 0.78 0.83 0.53 0.63 0.40 0.48 Belarus 0.53 0.59 0.70 0.74 0.56 0.65 0.33 0.39 Belgium 0.76 0.78 0.84 0.85 0.68 0.73 0.75 0.76 Belize 0.38 0.40 0.56 0.58 0.21 0.29 0.36 0.34 Benin 0.21 0.22 0.33 0.33 0.12 0.15 0.18 0.20 Bhutan 0.38 0.44 0.43 0.47 0.24 0.37 0.47 0.50 Bolivia 0.44 0.48 0.49 0.55 0.27 0.30 0.56 0.59 Bosnia and Herzegovina 0.55 0.60 0.64 0.68 0.42 0.47 0.59 0.64 Botswana 0.46 0.47 0.48 0.48 0.43 0.47 0.48 0.47 Brazil 0.65 0.68 0.65 0.68 0.55 0.55 0.77 0.82 Brunei Darussalam 0.57 0.63 0.66 0.69 0.45 0.53 0.61 0.66 Bulgaria 0.57 0.63 0.67 0.69 0.57 0.62 0.46 0.57 Burkina Faso 0.22 0.24 0.26 0.28 0.11 0.15 0.31 0.27 Burundi 0.23 0.26 0.30 0.31 0.02 0.06 0.37 0.42 Cabo Verde 0.37 0.43 0.45 0.49 0.39 0.43 0.28 0.38 Cambodia 0.36 0.40 0.35 0.41 0.30 0.39 0.43 0.39 Cameroon 0.27 0.30 0.23 0.28 0.13 0.15 0.45 0.46 Canada 0.67 0.69 0.78 0.79 0.64 0.69 0.58 0.60 Central African Republic 0.14 0.15 0.28 0.32 0.01 0.01 0.12 0.11 Chad 0.18 0.23 0.19 0.29 0.03 0.05 0.31 0.34 Chile 0.72 0.76 0.77 0.82 0.50 0.56 0.91 0.89 China 0.50 0.59 0.47 0.55 0.40 0.52 0.63 0.69 Colombia 0.61 0.64 0.64 0.67 0.42 0.48 0.76 0.76 Comoros 0.22 0.25 0.30 0.36 0.06 0.08 0.30 0.32 Congo, Dem. Rep. 0.19 0.21 0.14 0.17 0.06 0.05 0.38 0.40 Congo, Rep. 0.29 0.31 0.34 0.36 0.18 0.22 0.37 0.35 Costa Rica 0.61 0.66 0.65 0.68 0.54 0.68 0.62 0.63 Croatia 0.58 0.65 0.70 0.75 0.54 0.58 0.51 0.61 Cuba 0.21 0.24 0.24 0.29 0.07 0.12 0.31 0.30 Cyprus 0.62 0.68 0.76 0.82 0.68 0.77 0.42 0.44 Czech Republic 0.69 0.72 0.82 0.86 0.64 0.66 0.61 0.65 Denmark 0.78 0.79 0.93 0.92 0.88 0.90 0.52 0.56 Djibouti 0.28 0.30 0.45 0.47 0.06 0.09 0.34 0.33 (continued next page) 126 Information and Communications for Development 2018 Table DN.4 (continued) Digital Adoption Index Business subindex People subindex Government subindex Country 2014 2016 2014 2016 2014 2016 2014 2016 Dominica 0.45 0.50 0.54 0.55 0.37 0.47 0.44 0.48 Dominican Republic 0.46 0.50 0.51 0.52 0.30 0.37 0.56 0.60 Ecuador 0.52 0.57 0.66 0.68 0.33 0.39 0.58 0.63 Egypt, Arab Rep. 0.51 0.53 0.45 0.49 0.32 0.38 0.75 0.71 El Salvador 0.48 0.50 0.54 0.56 0.34 0.40 0.57 0.55 Equatorial Guinea 0.17 0.19 0.36 0.38 0.12 0.13 0.02 0.04 Estonia 0.77 0.83 0.76 0.85 0.73 0.80 0.81 0.85 Ethiopia 0.23 0.27 0.24 0.26 0.07 0.15 0.37 0.40 Fiji 0.43 0.46 0.52 0.54 0.30 0.38 0.47 0.47 Finland 0.79 0.81 0.92 0.92 0.84 0.83 0.61 0.67 France 0.74 0.75 0.76 0.77 0.69 0.73 0.78 0.76 Gabon 0.35 0.36 0.42 0.42 0.35 0.35 0.29 0.31 Gambia, The 0.33 0.36 0.38 0.40 0.22 0.29 0.39 0.39 Georgia 0.56 0.60 0.62 0.64 0.41 0.48 0.66 0.67 Germany 0.80 0.84 0.85 0.87 0.74 0.78 0.81 0.87 Ghana 0.38 0.45 0.35 0.42 0.27 0.39 0.51 0.55 Greece 0.58 0.61 0.69 0.71 0.61 0.68 0.43 0.42 Grenada 0.51 0.53 0.64 0.65 0.38 0.42 0.52 0.53 Guatemala 0.44 0.52 0.56 0.57 0.25 0.33 0.50 0.67 Guinea 0.21 0.21 0.20 0.13 0.10 0.15 0.31 0.34 Guinea-Bissau 0.23 0.26 0.26 0.30 0.07 0.10 0.35 0.38 Guyana 0.32 0.36 0.49 0.57 0.19 0.21 0.28 0.29 Haiti 0.25 0.25 0.34 0.31 0.11 0.12 0.30 0.32 Honduras 0.41 0.43 0.48 0.51 0.21 0.27 0.53 0.50 Hungary 0.64 0.69 0.67 0.77 0.61 0.65 0.63 0.65 Iceland 0.70 0.74 0.94 0.97 0.76 0.82 0.42 0.42 India 0.44 0.51 0.43 0.50 0.16 0.23 0.74 0.80 Indonesia 0.39 0.46 0.34 0.42 0.30 0.41 0.54 0.54 Iran, Islamic Rep. 0.42 0.51 0.42 0.53 0.26 0.44 0.56 0.55 Iraq 0.26 0.30 0.28 0.33 0.17 0.20 0.33 0.38 Ireland 0.64 0.66 0.81 0.83 0.62 0.65 0.49 0.50 Israel 0.75 0.79 0.73 0.77 0.67 0.74 0.85 0.85 Italy 0.73 0.77 0.73 0.75 0.64 0.68 0.83 0.87 Jamaica 0.44 0.50 0.52 0.58 0.33 0.42 0.48 0.49 Japan 0.82 0.83 0.73 0.76 0.79 0.84 0.93 0.91 Jordan 0.52 0.55 0.52 0.50 0.45 0.57 0.60 0.58 Kazakhstan 0.63 0.67 0.54 0.60 0.53 0.57 0.83 0.84 Kenya 0.40 0.45 0.51 0.57 0.15 0.20 0.55 0.59 Kiribati 0.20 0.21 0.41 0.40 0.04 0.09 0.15 0.15 Korea, Rep. 0.84 0.86 0.74 0.75 0.80 0.84 0.99 0.98 (continued next page) Data for Development Indicators 127 Table DN.4 (continued) Digital Adoption Index Business subindex People subindex Government subindex Country 2014 2016 2014 2016 2014 2016 2014 2016 Kuwait 0.63 0.63 0.71 0.73 0.71 0.67 0.48 0.50 Kyrgyz Republic 0.43 0.50 0.49 0.61 0.31 0.35 0.49 0.54 Lao PDR 0.20 0.26 0.24 0.34 0.14 0.17 0.22 0.27 Latvia 0.69 0.73 0.70 0.77 0.63 0.71 0.74 0.71 Lebanon 0.52 0.57 0.62 0.67 0.50 0.56 0.44 0.49 Lesotho 0.26 0.29 0.25 0.30 0.21 0.25 0.33 0.32 Liberia 0.21 0.24 0.27 0.29 0.10 0.13 0.24 0.29 Lithuania 0.75 0.79 0.77 0.80 0.67 0.75 0.80 0.83 Luxembourg 0.84 0.86 0.93 0.94 0.85 0.87 0.74 0.77 Macedonia, FYR 0.50 0.57 0.61 0.66 0.48 0.51 0.43 0.55 Madagascar 0.23 0.25 0.34 0.38 0.05 0.06 0.32 0.31 Malawi 0.24 0.26 0.36 0.39 0.04 0.07 0.30 0.32 Malaysia 0.65 0.69 0.52 0.55 0.59 0.64 0.85 0.87 Maldives 0.48 0.51 0.63 0.64 0.47 0.58 0.35 0.31 Mali 0.31 0.29 0.31 0.28 0.25 0.22 0.39 0.37 Malta 0.78 0.86 0.92 0.94 0.72 0.79 0.71 0.84 Marshall Islands 0.19 0.22 0.44 0.52 0.07 0.09 0.07 0.04 Mauritania 0.30 0.34 0.32 0.38 0.20 0.24 0.38 0.39 Mauritius 0.54 0.62 0.58 0.63 0.47 0.57 0.58 0.65 Mexico 0.54 0.60 0.59 0.63 0.35 0.44 0.68 0.74 Moldova 0.56 0.60 0.68 0.70 0.44 0.55 0.54 0.57 Mongolia 0.52 0.54 0.63 0.65 0.29 0.35 0.65 0.61 Montenegro 0.54 0.62 0.55 0.62 0.59 0.68 0.49 0.55 Morocco 0.52 0.56 0.54 0.60 0.40 0.42 0.63 0.64 Mozambique 0.28 0.25 0.33 0.26 0.13 0.17 0.37 0.33 Myanmar 0.17 0.26 0.22 0.28 0.11 0.27 0.18 0.23 Namibia 0.37 0.38 0.52 0.50 0.26 0.34 0.33 0.31 Nepal 0.30 0.37 0.32 0.35 0.16 0.25 0.42 0.50 Netherlands 0.83 0.84 0.92 0.91 0.75 0.80 0.81 0.81 New Zealand 0.67 0.71 0.76 0.77 0.72 0.79 0.53 0.56 Nicaragua 0.38 0.46 0.47 0.50 0.26 0.37 0.40 0.50 Niger 0.16 0.16 0.23 0.24 0.04 0.05 0.20 0.18 Nigeria 0.37 0.42 0.29 0.36 0.17 0.21 0.65 0.68 Norway 0.78 0.80 0.86 0.88 0.78 0.81 0.70 0.72 Oman 0.64 0.65 0.65 0.68 0.60 0.65 0.68 0.63 Pakistan 0.37 0.40 0.40 0.47 0.13 0.16 0.57 0.57 Panama 0.55 0.57 0.62 0.62 0.45 0.55 0.56 0.55 Papua New Guinea 0.31 0.34 0.53 0.55 0.06 0.09 0.32 0.38 Paraguay 0.46 0.54 0.58 0.62 0.30 0.37 0.51 0.64 Peru 0.52 0.55 0.59 0.61 0.34 0.43 0.62 0.62 (continued next page) 128 Information and Communications for Development 2018 Table DN.4 (continued) Digital Adoption Index Business subindex People subindex Government subindex Country 2014 2016 2014 2016 2014 2016 2014 2016 Philippines 0.44 0.49 0.53 0.57 0.38 0.44 0.41 0.47 Poland 0.65 0.69 0.73 0.76 0.64 0.68 0.58 0.63 Portugal 0.74 0.79 0.72 0.76 0.66 0.73 0.83 0.87 Qatar 0.68 0.71 0.74 0.76 0.70 0.76 0.60 0.60 Romania 0.62 0.64 0.64 0.65 0.51 0.57 0.71 0.72 Russian Federation 0.69 0.74 0.65 0.71 0.60 0.70 0.82 0.82 Rwanda 0.41 0.43 0.41 0.42 0.13 0.19 0.69 0.67 Samoa 0.31 0.36 0.46 0.51 0.17 0.25 0.30 0.33 Saudi Arabia 0.66 0.67 0.66 0.68 0.69 0.72 0.64 0.60 Senegal 0.33 0.35 0.39 0.39 0.19 0.23 0.41 0.43 Serbia 0.60 0.69 0.63 0.67 0.50 0.57 0.68 0.82 Seychelles 0.56 0.60 0.65 0.68 0.51 0.58 0.51 0.53 Sierra Leone 0.24 0.27 0.19 0.19 0.13 0.20 0.40 0.42 Singapore 0.87 0.87 0.84 0.85 0.80 0.80 0.96 0.96 Slovak Republic 0.65 0.69 0.69 0.75 0.59 0.67 0.66 0.64 Slovenia 0.64 0.71 0.83 0.86 0.59 0.63 0.51 0.65 Solomon Islands 0.21 0.27 0.32 0.42 0.12 0.15 0.20 0.24 South Africa 0.59 0.64 0.65 0.69 0.45 0.50 0.67 0.73 Spain 0.74 0.77 0.76 0.78 0.62 0.67 0.85 0.84 Sri Lanka 0.43 0.48 0.40 0.44 0.28 0.38 0.61 0.61 St. Kitts and Nevis 0.47 0.53 0.72 0.71 0.50 0.61 0.20 0.25 St. Lucia 0.43 0.40 0.55 0.44 0.35 0.35 0.40 0.41 St. Vincent and the Grenadines 0.46 0.50 0.65 0.66 0.33 0.40 0.38 0.43 Sudan 0.29 0.30 0.30 0.37 0.19 0.20 0.37 0.32 Suriname 0.43 0.49 0.48 0.52 0.55 0.62 0.27 0.32 Swaziland 0.28 0.32 0.44 0.46 0.17 0.20 0.24 0.28 Sweden 0.80 0.83 0.92 0.94 0.85 0.85 0.64 0.70 Switzerland 0.79 0.82 0.89 0.89 0.84 0.89 0.66 0.69 Syrian Arab Republic 0.27 0.32 0.44 0.51 0.18 0.22 0.17 0.22 Tajikistan 0.29 0.32 0.38 0.42 0.20 0.24 0.28 0.32 Tanzania 0.30 0.34 0.29 0.28 0.11 0.17 0.48 0.57 Thailand 0.57 0.62 0.55 0.57 0.57 0.68 0.58 0.62 Timor-Leste 0.26 0.29 0.23 0.27 0.26 0.29 0.30 0.31 Togo 0.21 0.25 0.37 0.37 0.09 0.14 0.17 0.24 Tonga 0.29 0.33 0.39 0.45 0.17 0.22 0.31 0.31 Trinidad and Tobago 0.51 0.59 0.57 0.64 0.49 0.57 0.48 0.55 Tunisia 0.53 0.56 0.60 0.61 0.41 0.46 0.56 0.60 Turkey 0.60 0.63 0.64 0.68 0.38 0.43 0.77 0.79 Turkmenistan 0.24 0.27 0.41 0.44 0.23 0.29 0.08 0.08 Tuvalu 0.25 0.29 0.50 0.53 0.17 0.26 0.07 0.07 (continued next page) Data for Development Indicators 129 Table DN.4 (continued) Digital Adoption Index Business subindex People subindex Government subindex Country 2014 2016 2014 2016 2014 2016 2014 2016 Uganda 0.28 0.34 0.28 0.32 0.10 0.14 0.45 0.56 Ukraine 0.45 0.54 0.61 0.67 0.38 0.47 0.37 0.47 United Arab Emirates 0.80 0.82 0.75 0.78 0.76 0.80 0.88 0.89 United Kingdom 0.74 0.76 0.88 0.90 0.77 0.80 0.55 0.59 United States 0.72 0.75 0.76 0.78 0.66 0.73 0.74 0.73 Uruguay 0.73 0.76 0.65 0.68 0.64 0.71 0.91 0.88 Uzbekistan 0.31 0.40 0.26 0.36 0.22 0.31 0.45 0.53 Vanuatu 0.27 0.32 0.43 0.51 0.14 0.20 0.23 0.26 Venezuela, RB 0.50 0.49 0.53 0.55 0.40 0.40 0.56 0.52 Vietnam 0.47 0.52 0.51 0.59 0.41 0.43 0.49 0.54 Yemen, Rep. 0.27 0.26 0.25 0.25 0.15 0.16 0.41 0.36 Zambia 0.29 0.34 0.30 0.33 0.14 0.18 0.44 0.52 Zimbabwe 0.30 0.33 0.38 0.43 0.17 0.21 0.36 0.35 All but a handful of countries increased their DAI scores scoring in the top 10 in both years for which data is available: in the period 2014–16, with most countries maintaining Austria, Germany, Japan, the Republic of Korea, Luxembourg, their relative positions. Overall, DAI was highly correlated Netherlands, Singapore, and Sweden (figure DN.2, panel a). with per capita income in 2014 and 2016, and there was All of the largest improvements came from middle-income improvement across the income spectrum (figure DN.1). countries, which as a group grew more than high-income The positive correlation between income and digital adop- countries. On the other hand, low-income countries grew tion is again apparent at the aggregate level. All of the top most slowly as a group, indicating a failure to converge with 10 countries in 2016 were high income, with eight of them the rest of the world (figure DN.2, panel b). Figure DN.1 Changes in Digital Adoption Index scores and per capita income, 2014–16 1.0 0.8 Digital Adoption Index scores 0.6 0.4 0.2 0 500 2,500 10,000 50,000 GDP per capita (US$, logarithmic scale) Source: World Bank 2018. 130 Information and Communications for Development 2018 Figure DN.2 Top Digital Adoption Index scores, 2016, and largest improvements, 2014–16 a. Scores, 2016 b. Change in scores, 2014–16 Top scores Largest improvements Singapore Iran, Islamic Rep. Luxembourg Uzbekistan Austria Guatemala Korea, Rep. Myanmar Malta Ukraine Germany China Netherlands Serbia Japan Ghana Estonia Nicaragua Sweden Paraguay Average scores Average improvement High income Lower-middle income Upper-middle income Upper-middle income World World Lower-middle income High income Low income Low income 0 0.2 0.4 0.6 0.8 1.0 0 0.02 0.04 0.06 0.08 0.10 Source: World Bank 2018. Notes 13. For the complete methodology of the DAI, see “Digital Adoption Index” at http://www.worldbank.org/en/publication 1. See “World Bank Country and Lending Groups” at https:// /wdr2016/Digital-Adoption-Index. datahelpdesk.worldbank.org/knowledgebase/articles/906519 14. The only indicator that is normalized within years is the -world-bank-country-and-lending-groups. Online Service Index calculated by the UN Department of 2. SDG Indicators Global Database at https://unstats.un.org Economic and Social Affairs for each of its biennial reports. /sdgs/indicators/database/?indicator=9.c.1. The latter does not make available source data that would 3. https://data.worldbank.org/indicator/it.net.user.zs. allow for normalizing the Online Service Index across years. 4. http://a4ai.org/broadband-pricing-data-2017/. 5. See “Broadband Portal” at http://www.oecd.org/sti/broadband /broadband-statistics/. References 6. https://researchictafrica.net/ramp_indices_portal/. ITU (International Telecommunications Union). 2017. Measuring 7. https://www.internetsociety.org/resources/doc/2017/sidsreport/. the Internet Society Report 2017. Vol. 2: ICT Country Profiles. Geneva: ITU. https://www.itu.int/en/ITU-D/Statistics/Pages 8. http://www.ictdata.org. /publications/mis2017.aspx. 9. See “Broadband Portal” at http://www.oecd.org/sti/broadband Mercier, Rémi. 2015. “How We Put Together a List of 2600+ Open /broadband-statistics/. Data Portals around the World for the Open Data Community.” 10. https://icdppc.org/participation-in-the-conference/list-of OpenDataSoft Blog, November 2. https://www.opendatasoft -accredited-members/. .com/2015/11/02/how-we-put-together-a-list-of-1600-open-data 11. http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D -portals-around-the-world-to-help-open-data-community/. -Legislation/eCom-Data-Protection-Laws.aspx. World Bank. 2018. Digital Adoption Index. 2018 update. http:// 12. Internet exchange point directory reports (http://wwww.pch www.worldbank.org/en/publication/wdr2016/Digital .net/ixpdir/summary). -Adoption-Index. Data for Development Indicators 131 Bibliography 3GPP. 2017. “Industry Support for 3GPP NR Announcement.” .netdna-cdn.com/wp-content/uploads/2016/05/Measur Press release, December 21. http://www.3gpp.org/news-events ingImpactsofMobileDataServices_ResearchBrief2.pdf. /3gpp-news/1931-industry_pr_5g. Andrews, D., C. Criscuolo, and P. N. Gal. 2016. “The Best versus Abazorius, Abby. 2016. “How Data Can Help Change the World.” the Rest: The Global Productivity Slowdown, Divergence MIT News, Cambridge, MA, September 26. http://news.mit across Firms and the Role of Public Policy.” Productivity .edu/2016/IDSS-celebration-big-data-change-world-0926. Working Paper 5, OECD Publishing, Paris. https://www.oecd Abdelkafi, Nizar, Sergiy Makhotin, and Thorsten Posselt. 2013. .org/global-forum-productivity/library/OECD%20Produc “Business Model Innovations for Electric Mobility: What tivity%20Working%20Paper%20N%C2%B05.pdf. Can Be Learned from Existing Business Model Patterns?” Anthony, Sebastian. 2012. “$1.5 Billion: The Cost of Cutting International Journal of Innovation Management 17 (1). London-Tokyo Latency by 60ms.” ExtremeTech, March 20. https://doi.org/10.1142/S1363919613400033. https://www.extremetech.com/extreme/122989-1-5-billion Accenture. 2015. “Guarding and Growing Personal Data Value.” -the-cost-of-cutting-london-toyko-latency-by-60ms. https://www.accenture.com/t20150821T065218__w__ /us-en APEC (Asia-Pacific Economic Cooperation). 2015. The APEC /_acnmedia/Accenture/Conversion-Assets/DotCom/Docu Privacy Framework. Singapore: APEC. ments/Global/PDF/Dualpub_15/Accenture-Guarding-and Armstrong, Mark. 2006. “Competition in Two-Sided Markets.” -Growing-Personal-Data-Value-Narrative-Repo. RAND Journal of Economics 17 (3): 688–91. Agrawal, Ravi. 2017. “India Supreme Court Rules Privacy a Badran, Mona Farid. 2014. “Access and Use of ICTs in Female- ‘Fundamental Right’ in Landmark Case.” CNN.com, Owned SMEs in Selected Arab Countries and Brazil: A August 24. https://www.cnn.com/2017/08/24/asia/indian Comparative Study.” Working Paper 2014/12, Maastricht -court-right-to-privacy/index.html. School of Management. https://www.msm.nl/resources Akamai. 2016. Q1 2016: State of the Internet. https://www /uploads/2014/05/MSM-WP2014-12.pdf. . akamai .com/us/en/our-thinking/state-of-the-internet-re Bain and Company. 2013. “Big Data: The Organizational Challenge.” port/index.jsp. https://www.bain.com/insights/big_data_the_organizational Aker, Jenny C. 2010. “Information from Markets Near and Far: _challenge/. Mobile Phones and Agricultural Markets in Niger.” American Bakhshi, Hasan, Albert Bravo-Biosca, and Juan Mateos-Garcia. Economic Journal: Applied Economics 2 (3): 46–59. https:// 2014. “The Analytical Firm: Estimating the Effect of Data and www.aeaweb.org/articles?id=10.1257/app.2.3.46. Online Analytics on Firm Performance.” Working Paper Alliance for Affordable Internet. 2016. “The Impacts of Emerging 14/05. https://www.nesta.org.uk/report/the-analytical-firm Mobile Data Services in Developing Countries.” Research -estimating-the-effect-of-data-and-online-analytics-on-firm Brief No. 2. http://1e8q3q16vyc81g8l3h3md6q5f5e.wpengine -performance/. 133 ———. 2014. “Inside the Datavores: Technical Report.” Nesta. Booth, Gregory D. 2017. “A Long Tail in the Digital Age: Music https://www.nesta.org.uk/report/inside-the-datavores-tech Commerce and the Mobile Platform in India.” Asia Music nical-report/. 48 (1): 85–113. Barnett, Jr., Thomas. 2016. “The Zettabyte Era Officially Begins Bort, Julie. 2016. “Now Facebook Plans to Eat the $500 Billion (How Much Is That?).” SP360: Service Provider, September 9. Telecom Equipment Market.” Business Insider, November 1. https://blogs.cisco.com/sp/the-zettabyte-era-officially-be http://www.businessinsider.com/facebook-voyager-optic gins-how-much-is-that. al-switch-telecom-infra-project-2016-11. BBC (British Broadcasting Corporation). 2017. “Google Apologises Bosco, Claudio, Victor Alegana, Tom Bird, Carla Pezzulo, Graeme after Ads Appear Next to Extremist Content.” March 20. Hornby, Alessandro Sorichetta, Jessica Steele, Cori https://www.bbc.com/news/business-39325916. Ruktanonchai, Nick Ruktanonchai, Erik Wetter, Linus ———. 2018. “Eric Schneiderman, New York Attorney General, Bengtsson, and Andrew J. Tatem. 2017. Mapping Indicators of Quits Amid Assault Reports.” BBC.com, May 8. https://www Female Welfare at High Spatial Resolution. WorldPop Project .bbc.com/news/world-us-canada-44035718. and Flowminder Foundation. http://www.data2x.org/wp Bean, Charles. 2016. “Independent Review of UK Economic -content /uploads/2017/02/Mapping-Indicators-of-Female Statistics.” Final Report of the Independent Review of UK -Welfare-at-High-Spatial-Resolution.pdf. Economic Statistics (to the Government of the United Kingdom), Brahm, Chris, Aaron Cheris, and Lori Sherer. 2016. “What Big led by Professor Sir Charles Bean of the London School of Data Means for Customer Loyalty.” Brief, Bain and Company, Economics. https://www.gov.uk/government/publi cations August 7. http://www.bain.com/publications/articles/what /independent-review-of-uk-economic-statistics-final-report. -big-data-means-for-customer-loyalty.aspx. Bell, M. 2015. “BMW, Audi, Daimler Buy Nokia’s Mapping Unit: Brandom, Russell. 2017. “Apple’s New Anti-tracking System Will An Autonomous Future Is Nigh.” Car, December 8. www Make Google and Facebook Even More Powerful.” The Verge, .carmagazine.co.uk/car-news/industry-news/mercedes-benz June 6. https://www.theverge.com/2017/6/6/15747300/apple /bmw-audi-and-daimler-purchase-nokias-here-system-an-au -safari-ad-tracking-cookie-blocker-google-facebook-privacy. tonomous-future-is-nigh/. Broadband Commision Working Group on Education. 2017. Bergvinson, David. 2017. “Digital Agriculture Empowers Farmers.” “Digital Skills for Life and Work.” http://unesdoc.unesco.org Business Today, January 15. https://www.businesstoday.in /images/0025/002590/259013e.pdf. /magazine/features/digital-agriculture-empowers-farmers Brynjolfsson, Erik, Lorin M. Hitt, and Heekyung Hellen Kim. /story/242966.html. 2011. “Strength in Numbers: How Does Data-Driven Decision Bershidsky, Leonard. 2017. “Google and Facebook Too Can Be Making Affect Firm Performance?” Unpublished. https:// Disrupted.” Bloomberg, December 8. https://www.bloomberg papers.ssrn.com/sol3/papers.cfm?abstract_id=1819486. .com/view/articles/2017-12-08/google-and-facebook-too BSA The Software Alliance. 2016. “Seizing Opportunity through -can-be-disrupted. License Compliance.” http://globalstudy.bsa.org/2016/index Beye, Birago Diene Moctar. 2015. “The Impact of OTT on .html. Telecommunication Services Provided by Operators in Burns, Thaddeus. 2017. “Regulating Machine Data: Less Is More Senegal.” Presentation at SAWAS 2015, Marrakesh, Morocco. for Global Growth.” WIPO: Magazine, December. http://www https://financedocbox.com/Tax_Planning/72673602-The .wipo.int/wipo_magazine/en/2017/06/article_0005.html. -impact-of-ott-on-telecommunication-services-provided -by-operators-in-senegal.html. Burson, Forrest. 2016. “6 Ways Amazon Is Changing Supply Chain Management in 2016.” Software Advice. https://www.soft Boateng, Richard, Joseph Budu, Alfred Sekyere Mbrokoh, Eric wareadvice.com/resources/amazon-supply-chain-manage Ansong, Sheena Lovia Boateng, and Augustus Barnnet ment/. Anderson. 2017. “Digital Enterprises in Africa: A Synthesis of Current Evidence.” http://believeoverhope.org/pearlrichards BusinessTech. 2018. “A First Look at Amazon’s New Offices in /index.php/2017/08/20/digital-enterprises-in-africa-a-syn Cape Town.” https://businesstech.co.za/news/business/227745 thesis-of-current-evidence/. /a-first-look-at-amazons-new-offices-in-cape-town/. Boffey, Daniel. 2017. “Google Fined Record €2.4bn by EU over Caldwelladr, Carole, and Emma Graham-Harrison. 2018. Search Engine Results.” The Guardian, June 27. https://www “Revealed: 50 Million Facebook Profiles Harvested for .theguardian.com/business/2017/jun/27/google-braces-for Cambridge Analytica in Major Data Breach.” The Guardian, -record-breaking-1bn-fine-from-eu. March 17. https://www.theguardian.com/news/2018/mar/17 Bohannon, John. 2016. “Satellite Images Can Map Poverty.” /cambridge-analytica-facebook-influence-us-election. Science, August 18. http://www.sciencemag.org/news/2016/08 Campolo, Alex, Madelyn Sanfilippo, Meredith Whittaker, and Kate /satellite-images-can-map-poverty. Crawford. 2017. “AI Now 2017 Report.” https://assets.content 134 Information and Communications for Development 2018 f u l . co m / 8 w p r h hv n p f c 0 / 1 A 9 c 3 Z TC Z a 2 K EY M 6 4 W s Research 43 (3): 345–54. https://msbfile03.usc.edu/digital c2a/8636557c5fb14f2b74b2be64c3ce0c78/_AI_Now measures/mayzlin/intellcont/chevalier_mayzlin06-1.pdf. _Institute_2017_Report_.pdf. CIGI (Centre for International Governance Innovation). 2017. Cannataci, Joseph. 2016. “A More In-Depth Look at Open Data “2017 CIGI-Ipsos Global Survey on Internet Security and and Big Data.” Annex 11 in Report of the Special Rapporteur on Trust.” https://www.cigionline.org/internet-survey-2017. the Right to Privacy. A/HRC/31/64. Geneva: Office of the High Cisco. 2017. “Cisco Visual Networking Index: Forecast and Commissioner for Human Rights. Methodology, 2016–2021.” https://www.cisco.com/c/en/us ———. 2017. “Report of the Special Rapporteur on the Right to /solutions/collateral/service-provider/visual-networking-in Privacy.” Doc. A/72/43103. Office of the High Commissioner dex-vni/complete-white-paper-c11-481360.pdf. for Human Rights, Geneva. CNNIC (China Internet Network Information Center). 2016. Capgemini Consulting. 2018. “Open Data Maturity in Europe, 2017.” “Statistical Report on Internet Development in China.” https://www.capgemini.com/wp-content/uploads/2017/11/rep https://cnnic.com.cn/IDR/ReportDownloads/. ort-open-data-maturity-in-europe-2017.pdf. Confessore, Nicholas, Gabriel J. X. Dance, Richard Harris, and Carey, Bjorn. 2016. “Smartphone Speech Recognition Can Write Mark Hansen. 2018. “The Follower Factory.” New York Times, Text Messages Three Times Faster than Human Typing.” January 27. https://www.nytimes.com/interactive/2018/01/27 Stanford News Services, August 24. https://news.stanford /technology/social-media-bots.html. . edu/2016/08/24/stanford-study-speech-recognition-fast Constantinou, Ioanna, Attila Marton, and Virpi Kristiina er-texting/. Tuunainen. 2017. “Four Models of Sharing Economy Casselman, Ben. 2017. “A Start-Up Slump Is a Drag on the Platforms.” MIS Quarterly Executive 16 (4): 231–51. http:// Economy; Big Business May Be to Blame.” New York Times, misqe.org/ojs2/index.php/misqe/article/viewFile/798/474. September 20. https://www.nytimes.com/2017/09/20/busi Constine, Josh. 2016. “Facebook Has Connected 40m People with ness/economy/startup-business.html. Internet.org.” Tech Crunch, November 2. https://techcrunch Caughill, P. 2017. “Skype Can Now Translate Your Voice Calls into .com/2016/11/02/omnipresent/. 10 Different Languages in Real-Time.” Futurism, April 10. Council of Europe. 1985. “Convention for the Protection of https://futurism.com/skype-can-now-translate-your-voice Individuals with Regard to Automatic Processing of Personal -calls-into-10-different-languages-in-real-time/. Data.” Council of Europe, Strasbourg. https://www.coe.int/en Cavoukian, Ann. 2017. “Ethical Standards in Artificial Intelligence: /web/conventions/full-list/-/conventions/treaty/108. AI Ethics by Design.” Presentation to UN Global Pulse, Cutler, Ben, Spencer Fowers, Jeffrey Kramer, and Eric Peterson. 2017. December 13. “Want an Energy-Efficient Data Center? Build It Underwater.” CGAP (Consultative Group to Assist the Poor). 2015. “Mobile IEEE Spectrum: Technology, Engineering, and Science News, Money and APIs.” World Bank, Washington, DC. http://www February 21. http://spectrum.ieee.org/computing/hardware .cgap.org/blog/series/mobile-money-and-apis. /want-an-energyefficient-data-center-build-it-underwater. Chander, Anupam, and Uyen Lê. 2015. “Data Nationalism.” Emory da Silva, Issa Sikiti. 2014. “Mali Digital Plan 2020 to Reorganise Law Journal 64 (3): 677–739. http://law.emory.edu/elj Economy.” Biztech Africa, December 9. http://www.biztechaf /content/volume-64/issue-3/articles/data-nationalism.html. rica.com/article/mali-digital-plan-2020-reorganiseecon Chang, Lulu. 2016. “The Latest Data Breach Involves the Voting omy/9327/. Records of 93.4 Million Mexican Citizens.” Digital Trends, Dallas, George. 2014. “Making Sense of Internet Platforms: April 23. https://www.digitaltrends.com/computing/mexico Network Effects and Two Sided Markets.” https://georgemdal -voting-breach/. las.wordpress.com/2014/06/05/making-sense-of-internet Change Dyslexia. 2018. “The Best Tool to Detect and Improve -platforms-network-effects-and-two-sided-markets/. Dyslexia Related Skills” (in Spanish). https://www.changedys Data for Climate Action. 2017. “Grand Prize Winner: Electro- lexia.org/. Mobility: Cleaning Mexico City’s Air with Transformational Cheng, Steve, Matthias Daub, Axel Domeyer, and Martin Climate Policies through Big Data Pattern Analysis in Traffic Lundqvist. 2017. “Using Blockchain to Improve Data and Social Mobility.” http://www.dataforclimateaction.org Management in the Public Sector.” McKinsey and Company. /meet-the-winners/. https://www.mckinsey.com/business-functions/digital-mc Data Privacy Project. 2018. “Code of Conduct: Data Science Code kinsey/our-insights/using-blockchain-to-improve-data of Professional Conduct.” http://www.datascienceassn.org -management-in-the-public-sector. /code-of-conduct.html. Chevalier, Judith, and Dina Mayzlin. 2006. “The Effect of Word of Dave, Paresh. 2017. “Credit Giant Equifax Says Social Security Mouth on Sales: Online Book Reviews.” Journal of Marketing Numbers, Birth Dates of 143 Million Consumers May Have Bibliography 135 Been Exposed.” Los Angeles Times, September 7. http://www world.com/article/3035595/emerging-technology/artificial .latimes.com/business/technology/la-fi-tn-equifax-data -intelligence-needs-your-data-all-of-it.html. -breach-20170907-story.html. EMC Corporation. 2014. “Digital Universe Invaded by Sensors.” Davenport, Thomas H., Don Cohen, and Al Jacobson. 2005. Press release, April 9. https://www.emc.com/about/news “Competing on Analytics.” Babson Executive Education: /press/2014/20140409-01.htm. Knowledge Research Center. http://www.babsonknowledge Enders, Albrecht, Harald Hungenberg, Hans-Peter Denker, and .org/analytics.pdf. Sebastian Mauch. 2008 “The Long Tail of Social Networking: Deloitte Review. 2013. “Data as the New Currency: Government’s Revenue Models of Social Networking Sites.” European Role in Facilitating the Exchange.” Deloitte Review 13. https:// Management Journal 26 (3): 199–211. deloitte.wsj.com/r iskandcompliance/files/2013/11 Ericsson. 2016. “Internet of Things to Overtake Mobile Phones by /DataCurrency_report.pdf. 2018: Ericsson Mobility Report.” Press release, June 1. https:// Dewey, Catherine. 2016. “98 Personal Data Points That Facebook www.ericsson.com/en/press-releases/2016/6/internet-of Uses to Target Ads to You.” Washington Post, August 18. -things-to-overtake-mobile-phones-by-2018-ericsson-mobil https://www.washingtonpost.com/news/the-intersect/wp ity-report. /2016/08/19/98-personal-data-points-that-facebook-uses-to Ernst and Young. 2016. “Digital Supply Chain: It Is All about That -target-ads-to-you/?utm_term=.8e8b6b4fb92b. Data.” https://www.ey.com/Publication/vwLUAssets/Digital Dhalla, Amira. 2016. “New Partnership with UN Women to Teach _ supply_chain_-_its_all_about_the_data/%24FILE / EY Key Digital Skills to Women.” Mozilla Learning. https://learn -digital-supply-chain-its-all-about-that-data-final.pdf. ing.mozilla.org/blog/new-partnership-with-un-women-to Espinet, X., W. Wang, and S. Mehndiratta. 2017. “Low-Budget -teach-key-digital-skills-to-women. Techniques for Road Network Mapping and Road Condition DLA Piper. 2018. “Data Privacy Scorebook.” https://www.dlapiper Assessment That Are Accessible to Transport Agencies in dataprotection.com/scorebox/index.html. Developing Countries.” Transportation Research Record: Journal of the Transportation Research Board no. 2634: 1–7. Dutch Data Center Association. 2017. “The Economic Impact of Multi-tenant Data Centers in the Netherlands.” https://www Etherington, Darrell. 2017. “Otonomo Raises $25m to Help .vijfhart.nl/wp-content/uploads/2017/02/report_-_2017 Automakers Make Money from Connected Cars.” TechCrunch, _-_economic_impact_dutch_data_centers.pdf. April 7. https://techcrunch.com/2017/04/07/otonomo-raises -25m-to-help-automakers-make-money-from-connected-cars/. Earth Institute of Colombia University and Ericsson. 2016. How EU Digital SME Alliance. 2017. “European Cybersecurity Strategy: Information and Communications Technology can Accelerate Fostering the SME Ecosystem.” https://www.cyberwatching Action on the Sustainable Development Goals. Ericsson. https:// .eu/news-events/news/european-cybersecurity-strategy-fos www.ericsson.com/res/docs/2016/ict-sdg.pdf. tering-sme-ecosystem. Economic Times. 2017. “Railways to Use Artificial Intelligence for European Commission. 2011. “Data Is the New Gold.” Opening Preventing Signal Failures.” November 22. https://energy Remarks, Press Conference on Open Data Strategy, December .economictimes.indiatimes.com/news/power/railways 12. https://ec.europa.eu/digital-single-market/en/news/data -to-use-artificial-intelligence-for-preventing-signal-fail -new-gold. ures/61747929. ———. 2018. “Communication from the Commission to the E-Estonia. 2017. “Artificial Intelligence Is the Next Step for European Parliament, the Council, the European Economic E-Governance in Estonia.” https://e-estonia.com/artificial-in and Social Committee and the Committee of the Regions telligence-is-the-next-step-for-e-governance-state-adviser ‘Towards A Common European Data Space.’” European -reveals/. Commission, Brussels. Einav, Liran, and Jonathan Levin. 2014. “The Data Revolution and ———. 2018. “New Measures to Boost Key Competences and Economic Analysis.” Innovation Policy and the Economy 14: 1–24. Digital Skills, as Well as the European Dimension of Eisenmann, Thomas R., Geoffrey Parker, and Marshall Van Education.” Press release, January 17. http://europa.eu/rapid Alstyne. 2011. “Platform Envelopment.” Strategic Management /press-release_IP-18-102_en.htm. Journal 32 (12): 1270–85. European Parliament. 2016. “Towards a Thriving Data-Driven Elahee, K., and S. Jugoo. 2013. “Ocean Thermal Energy for Economy.” http://www.europarl.europa.eu/sides/getDoc Air-Conditioning: Case Study of a Green Data Center.” Energy .do?pubRef=-//EP//TEXT+TA+P8-TA-2016-0089 +0+DOC Sources, Part A: Recovery, Utilization, and Environmental +XML+V0//EN. Effects 35 (7): 679–84. doi:10.1080/15567036.2010.504941. European Union. 2012. “Charter of Fundamental Rights of the Elgan, Mike. 2016. “Artificial Intelligence Needs Your Data, All of European Union.” 2012/C 326/02. http://eur-lex.europa.eu It.” ComputerWorld, February 22. https://www.computer /legal-content/EN/TXT/?uri=CELEX:12012P/TXT. 136 Information and Communications for Development 2018 ———. 2018. “The EU General Data Protection Regulation 374 (2083): 20160360. http://rsta.royalsocietypublishing.org (GDPR).” https://www.eugdpr.org. /content/roypta/374/2083/20160360.full.pdf. ———. n.d. “Data Protection Factsheet.” http://ec.europa.eu Ford. 2016. “From Autonomy to Snowtonomy: How Ford Fusion / justice/data-protection/files/factsheets/factsheet_data Hybrid Autonomous Research Vehicle Can Navigate in _protection_en.pdf. Winter.” https://media.ford.com/content/fordmedia/fna/us ———. n.d. “Data Protection: Rules for the Protection of Personal / en/news/2016/03/10/how-fusion-hybrid-autonomous Data Inside and outside the EU.” https://ec.europa.eu/info -vehicle-can-navigate-in-winter.html. /law/law-topic/data-protection_en. ———. 2016. “Ford Tripling Autonomous Vehicle Development Evans, David S. 2013. “Economics of Vertical Restraints for Multi- Fleet, Accelerating On-Road Testing of Sensors and Software.” sided Platforms.” Working Paper, Coase-Sandor Institute for https://media.ford.com/content/fordmedia/fna/us/en Law and Economics, Chicago. https://chicagounbound.uchi / news/2016/01/05/ford-tripling-autonomous-vehicle-de cago.edu/cgi/viewcontent.cgi?article=1187&context=law velopment-fleet--accelerating.html. _and_economics. Frieden, Rob. 2017. “The Internet of Platforms and Two-Sided ———. 2013. “Some Empirical Aspects of Multi-sided Platform Markets: Legal and Regulatory Implications for Competition Industries.” Review of Network Economics 2 (3): 191–209. and Consumers.” Working paper, College of Communications, Evans, David S., and Richard Schmalansee. 2016. Matchmakers: Pennsylvania State University, State College, PA. https://ssrn The New Economics of Multisided Platforms. Boston: Harvard .com/abstract=3051766. Business Review Press. Gaba, K. M., and B. Sánchez-Andrade Nuño. 2016. “Eyes in the Sky Facebook. 2017. “Global Government Request Report.” April 27. Help Track Rural Electrification.” World Bank blog, January https://transparency.facebook.com/. 21. http://blogs.worldbank.org/energy/eyes-sky-help-track -rural-electrification. Facebook: Internet.org. 2015. “One Year In: Internet.org Free Basic Services.” Blog, July 27. https://info.internet.org/en/blog/2015 Garrahan, Matthew. 2016. “Advertising: Facebook and Google /07/27/one-year-in-internet-org-free-basic-services/. Build a Duopoly.” Financial Times, June 23. https://www.ft .com/content/6c6b74a4-3920-11e6-9a05-82a9b15a8ee7. Facebook: Internet.org. n.d. “Where We’ve Launched.” https://info .internet.org/en/story/where-weve-launched/. Garret, R. Kelly. 2016. “Facebook’s Problem Is More Complicated Than Fake News.” Scientific American, November 17. https:// Fell, Gregory, and Mike Barlow. 2016. “Not All Data Is Created www.scientificamerican.com/article/facebook-s-problem-is Equal.” O’Reilly, April 25. https://www.oreilly.com/learning -more-complicated-than-fake-news/. /not-all-data-is-created-equal. Gartner. 2015. “Gartner Says Australian Organizations to Spend Fernandez, Liliana, and Richard Heeks. 2016. “Measuring the A$2.5 Billion on Data Center Systems in 2015.” Press release, Barriers to Big Data for Development: Design-Reality Gap May 17. http://www.gartner.com/newsroom/id/3054918. Analysis in Colombia’s Public Sector.” Working Paper, University of Manchester, Manchester, UK. GfK. 2017. “Willingness to Share Personal Data in Exchange for Benefits or Rewards.” https://www.gfk.com/fileadmin/user Fiegerman, Seth. 2018. “More than 20 States Sue to Stop FCC’s Net _ upload/countr y_one_pager/NL/images/Global-GfK Neutrality Repeal.” CNN, January 16. http://money.cnn _onderzoek_-_delen_van_persoonlijke_data.pdf. .com/2018/01/16/technology/net-neutrality-lawsuit/index .html. Global Fishing Watch. 2018. “Vessel Monitoring Data: Indonesia.” http://globalfishingwatch.org/indonesia-vms. Field, Matthew. 2018. “DuckDuckGo: The Private Search Engine Standing Up to Google.” Technology Intelligence, April 23. Goldstein, Phil. 2013. “Report: Skype Makes Up One-Third of All https://www.telegraph.co.uk/technology/2018/04/23/duck International Phone Traffic.” FierceWireless, February 15. duckgo-tiny-private-search-engine-standing-google/. https://www.fiercewireless.com/wireless/report-skype-makes FlashGlobal. 2018. “Walmart: 3 Keys to Successful Supply Chain -up-one-third-all-international-phone-traffic. Management Any Business Can Follow.” April 12. https:// Good Things Foundation. 2017. “Project: Reboot UK.” https:// flashglobal.com/blog/supply-chain-management-walmart/. www.goodthingsfoundation.org/projects/reboot-uk. Flexenclosure. 2017. “Flexenclosure to Deploy Two eCentre Government of Canada. 2017. “Policy on Information Cable Landing Stations for SSCC in Samoa.” Press release, Management.” http://www.tbs-sct.gc.ca/pol/doc-eng.aspx?id February 27. http://www.flexenclosure.com/flexenclosure-de =12742. ploys-two -ecentre-cable-landing-stations-for-sscc-in Government of Colombia. 2018. “Emprende con Datos Finalizó -samoa/. con Más 60 Equipos Apoyados” [Start with Data Finished Floridi, Luciano, and Mariarosaria Taddeo. 2016. “What Is Data with More Than 60 Teams Supported]. Ministry of ICT. Ethics?” Philosophical Transactions of the Royal Society A http://www.mintic.gov.co/portal/604/w3-article-73884.html. Bibliography 137 Government of New Zealand. 2011. “New Zealand Data and Heeks, Richard, and Jaco Renken. 2016. “Data Justice for Information Management Principles.” Wellington. https://www Development: What Would It Mean?” Working paper, .ict.govt . nz/guidance-and-resources/open-government University of Manchester, Manchester, UK. /new-zealand-data-and-information-management-principles/. Hellemans, Alexander. 2015. “Why IoT Needs 5G.” IEEE Spectrum, Government of the United Kingdom. 2013. “Shakespeare Review: May 20. http://spectrum.ieee.org/tech-talk/computing An Independent Review of Public-Sector Information.” /networks/5g-taking-stock. May 15. https://www.gov.uk/government/publications/shake Henten, Anders, and Iwona Windekilde. 2016. “Transaction Costs speare-review-of-public-sector-information. and the Sharing Economy.” Info 18 (1): 1–15. ———. 2016. “Review of Information Security at HM Revenue Hesseldahl, Arik. 2015. “Why the Federal Government Sucks at and Customs.” http://webarchive.nationalarchives.gov.uk/20 Cyber Security.” Recode, June 23. https://www.recode.net 100 407 163917/http://www.hm-treasury.gov.uk/d/poynter /2015/6/23/11563798/why-the-federal-government -sucks _review250608.pdf. -at-cybersecurity. ———. 2017. “Digital Economy Act.” Hsu, Tiffany. 2018. “For Many Facebook Users, a ‘Last Straw’ That ———. n.d. “Data Sharing in Government.” Blog. http://datashar Led Them to Quit.” New York Times, March 21. https://www ing.org.uk/index.html. .nytimes.com/2018/03/21/technology/users-abandon-face Graef, Inge, Sih Yuliana Wahyuningtyas, and Peggy Valcke. 2015. book.html. “Assessing Data Access Issues in Online Platforms.” Hungenberg, Harald, Hans-Peter Denker, and Sebastian Mauch. Telecommunications Policy 39 (5): 375–87. 2008. “The Long Tail of Social Networking: Revenue Models Green, Andy. 2018. “Canada’s PIPEDA Breach Notification of Social Networking Sites.” European Management Journal Regulations Are Finalized!” Varonis, May 3. https://blog 26 (3): 199–211. . varonis.com/canadas-pipeda-breach-notification-regula IBM (International Business Machines Corporation). 2013. tions-finalized/. Harness the Power of Big Data: The IBM Big Data Platform. Greengard, Samuel. 2010. “Cloud Computing and Developing New York: McGraw-Hill. Nations.” Communications of the ACM 53 (5): 18–20. https:// ———. 2016. “IBM: 10 Key Marketing Trends for 2017 and Ideas cacm.acm.org/magazines/2010/5/87255-cloud-comput for Exceeding Customer Expectations.” https://www-01.ibm. ing-and-developing-nations/fulltext. com/common/ssi/cgi-bin/ssialias?htmlfid=WRL12345USEN. GSMA. 2017. Embracing the Digital Revolution Policies for Building ———. 2016. “What Is Big Data?” https://www.ibm.com/big-data the Digital Economy. Digital Transformation Report 2017. /us/en/. GSMA. https://www.gsma.com/publicpolicy/wp-content / uploads/2017/02/GSMA_DigitalTransformation Report ———. 2018. “IBM Watson for Genomics Helps Doctors Give 2017_Web.pdf. Patients New Hope.” https://www.ibm.com/watson/health /oncology-and-genomics/genomics/. Gunawan, Imana. 2017. “App Helps Indonesian Capital Get ‘Smart’ to Improve Public Services.” Humanosphere, January 6. http:// IDC (International Data Corporation). 2017. “Impact Assessment www.humanosphere.org/social-business/2017/01/app-helps of Odine Program: Open Data Incubator Europe.” https:// -indonesian-capital-get-smart-to-improve-public-services/. opendataincubator.eu/. Haddad, Ryan, Tim Kelly, Teemu Leinonen, and Vesa Saarinen. IEEE (Institute of Electrical and Electronics Engineers). 2016. 2014. Using Location Data from Mobile Phones to Enhance the “The IEEE Global Initiative on Ethics of Autonomous and Science of Delivery. Washington, DC: World Bank. https:// Intelligent Systems: Ethically Aligned Design; A Vision for openknowledge.worldbank.org/handle/10986/19316. Prioritizing Wellbeing with Artificial Intelligence and Harari, Y. N. 2017. Homo Deus: A Brief History of Tomorrow. Autonomous Systems V.1.” The IEEE Global Initiative on New York: Harper. Ethics of Autonomous and Intelligent Systems. http://stan dards.ieee.org/develop/indconn/ec/autonomous_systems Harwell, Drew. 2018. “Companies Race to Gather a Newly Prized .html. Currency: Our Body Measurements.” Washington Post, January 16. https://www.washingtonpost.com/business/econ Iimi, Atsushi, A. K. Farhad Ahmed, Edward Charles Anderson, omy/companies-race-to-gather-a-newly-prized-currency Adam Stone Diehl, Laban Mayo, Tatiana Peralta Quiros, and -our-body-measurements/2018/01/16/5af28d98-f6e8-11e7 Kulwinder Singh Rao. 2016. “New Rural Access Index: Main -beb6-c8d48830c54d_story.html?noredirect=on&utm Determinants and Correlation to Poverty.” Policy Research _term=.4eb7f0e18af4. Working Paper 7876, World Bank, Washington, DC. Heeks, Richard. 2017. “Development Informatics.” Working paper, Ingram, Matthew. 2017. “Facebook and Google Need to Be Global Development Institute, University of Manchester, Regulated, Says British News Industry.” Fortune, March 9. Manchester, UK. http://fortune.com/2017/03/09/facebook-google-regulated/. 138 Information and Communications for Development 2018 Intel. 2013. “Intel: She Will Connect Initiative.” https://www.intel Kastrenakes, Jacob. 2017. “The FCC Just Killed Net Neutrality.” The .com/content/www/us/en/corporate-responsibility/social-im Verge, December 14. https://www.theverge.com/2017/12/14 pact-and-educational-initiatives/she-will-connect.html. /16776154/fcc-net-neutrality-vote-results-rules-repealed. Internet Society. 2017. “A Case Study in Local Content Hosting: Katz, Raul, and P. Koutroumpis. 2012. “The Economic Impact Speed, Visits, and Cost of Access.” https://www.internetsoci of Telecommunications in Senegal.” Communications and ety.org/doc/case-study-local-content-rwanda. Strategies 2 (86): 21–42. https://www8.gsb.columbia.edu/citi IOM (International Organization for Migration). 2017. Fatal /sites/citi/files/files/Senegal_Telecoms_Report_version%20 Journeys: Improving Data on Missing Migrants. Part 1, vol. 3. finale%5B1%5D.pdf. Geneva: IOM. https://publications.iom.int/system/files/pdf Kelly, Heather, and Scott McLean. 2017. “Your Browser History Is /fatal_journeys_volume_3_part_1.pdf. for Sale, Here’s What You Need to Know.” CNNMoney, April 6. ITIF (Information Technology and Innovation Foundation). 2017. http://money.cnn.com/2017/04/05/technology/online “Cross-Border Data Flows: Where Are the Barriers, and What -privacy-faq/. Do They Cost?” May 1. https://itif.org/publications/2017/05 Kelly, Michael, and David Satola. 2017. “The Right to Be Forgotten.” /01/cross-border-data-flows-where-are-barriers-and-what University of Illinois Law Review 1. https://papers.ssrn.com -do-they-cost. /sol3/papers.cfm?abstract_id=2965685. ITU (International Telecommunication Union). 2015. Kelly, Tim, and Carlo Maria Rossotto. 2012. Broadband Strategies “Interoperability in the Digital Ecosystem.” GSR Discussion Handbook. Washington, DC: World Bank. https://openknow Paper, ITU, Geneva. https://www.itu.int/en/ITU-D ledge.worldbank.org/handle/10986/6009. /Conferences/GSR/Documents/GSR2015/Discussion_papers Kenney, M., and J. Zysman. 2016. “The Rise of the Platform _and_Presentations/Discussionpaper_interoperability.pdf. Economy.” Issues in Science and Technology 32 (3): 61–69. ———. 2017. “AI for Good Global Summit 2017.” https://www Kharpal, Arjun. 2016. “5 Reasons Why Uber Sold Its China .itu.int/en/ITU-T/AI/Pages/201706-default.aspx Business to Didi Chuxing.” CNBC, August 1. https://www ———. 2017. “ICT Facts and Figures.” https://www.itu.int/en .cnbc.com/2016/08/01/5-reasons-why-uber-sold-its-china /ITU-D/Statistics/Pages/facts/default.aspx. -business-to-didi-chuxing.html. ———. 2017. Measuring the Internet Society Report 2017. Vol. 2: Kitchin, Rob. 2014. The Data Revolution: Big Data, Open Data, ICT Country Profiles. Geneva: ITU. https://www.itu.int/en Data Infrastructures and Their Consequences. London: Sage. /ITU-D/Statistics/Pages/publications/mis2017.aspx. Koçoğlu, İpek, Salih Zeki İmamoğlu, Hüseyin İnce, and Halit Jean, Neal, Marshall Burke, Michael W. Xie, Matthew Davis, David Keskin. 2011. “The Effect of Supply Chain Integration B. Lobell, and Stegano Ermon. n.d. “Sustainability and on Information Sharing: Enhancing the Supply Chain Artificial Intelligence Lab: Combining Satellite Imagery and Performance.” Procedia—Social and Behavioral Sciences 24: Machine Learning to Predict Poverty.” Stanford University. 1630–49. https://www.sciencedirect.com/science/article/pii http://sustain.stanford.edu/predicting-poverty. /S1877042811015448. Juniper Research. 2016. “Understanding the Personal Data Krambeck, H., and L. Qu. 2015. “Toward an Open Transit Service Economy: The Emergence of a New Data Value-Exchange.” Data Standard in Developing Asian Countries.” Transportation https://mobileecosystemforum.com/personal-data-econ Research Record: Journal of the Transportation Research Board omy-whitepaper/. no. 2538: 30–36. Kalia, Amul. 2017. “Here’s How to Protect Your Privacy from Your Lal Das, Prasanna, Stephan Beisswenger, Srikanth Mangalam, and Internet Service Provider.” Electronic Frontier Foundation, Rasit Yuce. 2017. “Internet of Things: The New Government April 3. https://www.eff.org/deeplinks/2017/04/heres-how to Business Platform—A Review of Opportunities, Practices, -protect-your-privacy-your-internet-service-provider. and Challenges.” World Bank, Washington, DC. http://docu Kaminski, Piotr, Chris Rezek, Wolf Richter, and Marc Sorel. 2017. ments.worldbank.org/curated/en/610081509689089303/Inter “Protecting Your Critical Digital Assets: Not All Systems and net-of-things-the-new-government-to-business-platform-a Data Are Created Equal.” McKinsey and Company. https:// -review-of-opportunities-practices-and-challenges. www.mckinsey.com/business-functions/risk/our-insights Le Bras, Tom. 2015. “Online Overload—It’s Worse Than You /protecting-your-critical-digital-assets-not-all-systems-and Thought.” Infographic, Dash Lane Blog, July 21. https://blog -data-are-created-equal. .dashlane.com/infographic-online-overload-its-worse-than Kapoor, S. 2017. “Artificial Intelligence: Re-imagining Big Data’s -you-thought/. Applicability.” CXOtoday.com, July 17. www.cxotoday.com Leavitt, Neal. 2010. “Network-Usage Changes Push Internet Traffic /story /artificial-intelligence-re-imagining-big-datas-applic to the Edge.” Computer [IEEE] 43 (10): 13–15. http://ieeex ability/. plore.ieee.org/document/5604156. Bibliography 139 Leber, Jessica. 2013. “In a Data Deluge, Companies Seek to Fill a Magna Global. 2017. “Advertising Forecasts: Winter Update (Dec 4, New Role.” MIT Technology Review, May 22. https://www 2017).” https://www.magnaglobal.com/wp-content/uploads .technologyreview.com/s/513866/in-a-data-deluge-compan /2017/12/MAGNA-Global-Forecast_Winter-Update_Final.pdf. ies-seek-to-fill-a-new-role/. Marshall, Aarian. 2017. “The Secret Uber Data That Could Fix Lee, T. B. 2017. “Bitcoin’s Insane Energy Consumption, Explained.” Your Daily Commute.” Wired, February 3. https://www.wired https://arstechnica.com/tech-policy/2017/12/bitcoins-insane .com/2017/02/ubers-coughing-data-nyc-fix-commute/ -energy-consumption-explained/. Mastercard. 2012. “Nigerians Enjoy the Convenience, Pricing and Lehtiniemi, Tuukka. 2017. “Personal Data Spaces: An Intervention Choice of Online Shopping—MasterCard Survey.” Press in Surveillance Capitalism?” Surveillance and Society 15 (5). release, May 1. https://www1.mastercard.com/content/intelli https://ojs.library.queensu.ca/index.php/surveillance-and gence/en/research/press-release/2012/Nigerians-enjoy-the -society/article/view/6424. -convenience-pricing-and-choice-of-online-shopping.html. Lewis, Paul. 2017. “‘Our Minds Can Be Hijacked’: The Tech Matheson, Rob. 2016. “Optimizing Parking in the Busiest Places.” Insiders Who Fear a Smartphone Dystopia.” The Guardian, MIT News, October 25. http://news.mit.edu/2016/smarking October 6. https://www.theguardian.com/technology/2017 -data-analytics-optimizing-parking-1012. /oct/05/smartphone-addiction-silicon-valley-dystopia. Matthee, Marianne, and André Heymans. 2013. “How South Li, A. 2016. “Google Translate Adds 13 New Languages Bringing African SMEs Can Become Better Candidates for Export Total to 103.” https://9to5google.com/2016/02/17/google Finance.” Managing Global Transitions 11 (4): 391–407. -translate-13-languages. Matthey, James. 2018. “There’s Something Unusual about Simona Lindstedt, Catharina, and Daniel Naurin. 2005. “Transparency Halep’s Australian Open Dress.” News.com.au, January 26. and Corruption: The Conditional Significance of a Free https://www.news.com.au/sport/tennis/australian-open/theres Press.” Quality of Government Institute Working Paper -something-unusual-about-simona-haleps-australian -open 2005:5, University of Gothenburg, Sweden. https://www.qog -dress/news-story/28d7ce2da03fb00a9ad2bd5a79093156. .pol.gu . se/digitalAssets/1350/1350633_2005_5-lindstedt Mbiti, Isaac M., and David N. Weil. 2011. “Mobile Banking: The _naurin.pdf. Impact of M-Pesa in Kenya.” Working Paper 17129, National LinkedIn. 2017. “Brilent, the Intelligent Candidate Recommendation Bureau of Economic Research, Cambridge, MA. https://ssrn Engine.” https://www.linkedin.com/pulse/brilent-intelligent .com/abstract=1866089. -candidate-recommendation-engine-edouard-murat/. McGoogan, Cara. 2017. “Yahoo Hack Warning: What Happened Lomas, Natasha. 2017. “Form an Orderly Queue! Google Wants and Should You Be Worried?” The Telegraph, February 16. Your Blood (and Other Bodily Fluids). Oh and Your Medical https://www.telegraph.co.uk/technology/2017/02/16/yahoo Records.” TechCrunch. https://techcrunch.com/2017/04/21 -hack-warning-happened-should-worried/. /form-an-orderly-queue-google-wants-your-blood-and-oth McGregor, Lorna. 2017. “Ethical Development of AI.” Statement at er-bodily-fluids-oh-. the ITU AI for Good Summit, June. Luleå University of Technology. 2014. “National Agenda for the McKinsey Global Institute. 2011. Big Data: The Next Frontier for Internet of Things: Summary of the Project IoT Sweden.” Innovation, Competition, and Productivity. https://www https://iotsverige.se/wp-content/uploads/2014/12/Agendan .mckinsey.com/~/media/McKinsey/Business%20Functions -eng-29-sep-2014-korr4.pdf. / McKinsey%20Digital/Our%20Insights/Big%20data%20 Lyons, Elizabeth, Zakkoyya Lewis, Brian Mayrsohn, and Jennifer The%20next%20frontier%20for%20innovation/MGI_big Rowland. 2014. “Behavior Change Techniques Implemented _data_full_report.ashx. in Electronic Lifestyle Activity Monitors: A Systematic Content ———. 2016. The Age of Analytics: Competing in a Data-Driven Analysis.” Journal of Medical Internet Research 16 (8). http:// World. London: McKinsey Global Institute. https://www doi.org/10.2196/jmir.3469. .mckinsey.com/business-functions/mckinsey-analytics Maaroof, Abbas. 2015. “Big Data and the 2030 Agenda for /our-insights/the-age-of-analytics-competing-in-a-data Sustainable Development.” Report for UN-ESCAP. http:// -driven-world. www.unescap.org/sites/default/files/Final%20Draft_%20 ———. 2016. “Digital Globalization: The New Era of Global stock-taking%20report_For%20Comment_301115.pdf. Flows.” New York. https://www.mckinsey.com/business-func Madrigal, Alexis C. 2012. “I’m Being Followed: How Google—and tions/digital-mckinsey/our-insights/digital-globalization-the 104 Other Companies—Are Tracking Me on the Web.” The -new-era-of-global-flows. Atlantic, February 29. https://www.theatlantic.com/technol ———. 2017. “Digital Transformation: Raising Supply-Chain ogy/archive/2012/02/im-being-followed-how-google-151 Performance to New Levels.” https://www.mckinsey.com -and-104-other-companies-151-are-tracking-me-on-the / business-functions/operations/our-insights/digital-trans -web/253758/. formation-raising-supply-chain-performance-to-new-levels. 140 Information and Communications for Development 2018 ———. 2017. “Jobs Lost, Jobs Gained: Workforce Transitions in a http://www.datacenterdynamics.com/content-tracks/colo Time of Automation.” https://www.mckinsey.com/~/media -cloud/the-kingdom-of-bhutan-opens-first-government / Mc K i n s e y / G l o b a l % 2 0 T h e m e s / F u t u r e % 2 0 o f % 2 0 -data-center/98052.fullarticle. Organizations/What%20the%20future%20of%20work%20 Mozilla. 2018. “The Internet Health Report.” https://internet will%20mean%20for%20jobs%20skills%20and%20wages healthreport.org/. /MGI-Jobs-Lost-Jobs-Gained-Report-December-6-2017.ashx. Munster, Brett. 2017. “Could Blockchain Disrupt Facebook and ———. 2018. “Digital Reinvention: Unlocking the ‘How.’” https:// Google’s Business Model?” Medium. https://medium.com www.mckinsey.com/~/media/McKinsey/Business%20 / road-less-ventured/could-blockchain-disrupt-facebook Fu n c t i o n s / Mc Ki n s e y % 2 0 D i g i t a l / O u r % 2 0 I n s i g h t s -and-googles-business-model-fda614de492d. / Digital%20Reinvention%20Unlocking%20the%20how Murdock, Jason. 2018. “Mark Zuckerberg Says ‘Delete Facebook’ /Digital-Reinvention_Unlocking-the-how.ashx. Protest Had No Meaningful Impact on His Business.” Meeker, Mary. 2016. Internet Trends 2016: Code Conference. Menlo Newsweek, April 5. http://www.newsweek.com/zuckerberg Park, CA: Kleiner, Perkins, Caufield, Byers (KPCB). -says-deleting-facebook-has-no-meaningful-impact -his ———. 2017. Internet Trends 2017: Code Conference. Annual Report. -business-872876. Menlo Park, CA: Kleiner, Perkins, Caufield, Byers (KPCB). Naone, Erica. 2011. “How Useful Is Personalized Search?” MIT https://www.kleinerperkins.com/perspectives/internet Technology Review, April 11. https://www.technologyreview -trends-report-2017 .com/s/423596/how-useful-is-personalized-search/. Mercier, Rémi. 2015. “How We Put Together a List of 2600+ Open Neal, Dave. 2016. “Swedish Deseat Software Will Help You Delete Data Portals around the World for the Open Data Community.” Yourself from the Internet.” The Inquirer, November 25. OpenDataSoft Blog, November 2. https://www.opendatasoft https://www.theinquirer.net/inquirer/news/2478419/swedish .com/2015/11/02/how-we-put-together-a-list-of-1600-open -deseat-software-will-help-you-delete-yourself-from-the -data-por tals-around-the-world-to-help-open-data -internet. -community/. Nelson, P. 2016. “Just One Autonomous Car Will Use 4,000 GB of Microsoft Translator. n.d. “What Is Neural Network Based Data/Day.” Networkworld, December 7. https://www.network Translation?” https://microsofttranslator.uservoice.com world.com/article/3147892/internet/one-autonomous-car /knowledgebase/articles/1099027-what-is-neural-network -will-use-4000-gb-of-dataday.html. -based-translation Netcraft. 2017. “February 2017 Web Server Survey.” https://news .netcraft.com/archives/2017/02/27/february-2017-web-serv Millicom. 2016. “Tigo Business Paraguay’s Data Center Awarded er-survey.html. ‘Best Modular Implementation’ by Prestigious International Sector Magazine.” Press release, September 28. http://www Noack, Rick. 2018. “European Union’s New Privacy Law Made .millicom.com/media/6762386/Data-Center-Award_-Eng.pdf. Some Websites Go Dark Today; Here’s What Else Has Changed.” Los Angeles Times, May 25. http://www.latimes ———. n.d. “Investing in State-of-the-Art Data Centers.” News . com/business/technology/la-fi-tn-gdpr-europe-privacy feature. http://www.millicom.com/media/millicom-news -20180525-story.html. -features/data-centers. Nugraha, Y., Kautsarina, and A. S. Sastrosubroto. 2015. “Towards Mineo, Liz. 2017. “On Internet Privacy, Be Very Afraid.” Harvard Data Sovereignty in Cyberspace.” Working paper presented at Gazette, August 24. https://news.harvard.edu/gazette the Third International Conference on Information and /story/2017/08/when-it-comes-to-internet-privacy-be-very Communication Technology. https://www.cs.ox.ac.uk/files -afraid-analyst-suggests/. /7463/Towards%20Data%20Sovereig nit y%20in%20 MIT (Massachusetts Institute of Technology). “MIT Institute for Cyberspace_Nugraha.pdf. Data, Systems, and Society.” https://idss.mit.edu/. NUSDeltares. 2014. “Neptune OMS for Singapore Coastal Quality.” MIT (Massachusetts Institute of Technology) Technology Review. http://www.nusdeltares.org/projects/neptune/. 2014. “Researchers Test Personal Data Market to Find Out O’Brien, Sarah Ashley. 2016. “Why Facebook Was Launching a How Much Your Information Is Worth.” July 9. https://www Satellite into Space.” CNN.com, September 1. http://money .technologyreview.com/s/528866/researchers-test-personal-data .cnn.com/2016/09/01/technology/facebook-satellite-explo -market-to-find-out-how-much-your-information-is-worth/. sion/. Morey, Timothy, Theodore Forbath, and Allison Schoop. 2015. O’Conner, M. C. 2017. “Dumpster Diving Robots: Using AI for “Customer Data: Designing for Transparency and Trust.” Smart Recycling.” IQ, July 5. https://iq.intel.com/dumpster Harvard Business Review, May. -diving-robots-using-ai-for-smart-recycling/. Moss, Sebastian. 2017. “The Kingdom of Bhutan Opens First ODIHQ (Open Data Institute). 2015. “The Economic Impact of Government Data Center.” Data Centre Dynamics, March 27. Open Data: What Do We Already Know?” Medium, November 2. Bibliography 141 https://medium.com/@ODIHQ/the-economic-impact-of http://www.oecd-ilibrary.org/development /a-road-map-for -open-data-what-do-we-already-know-1a119c1958a0. -a-country-led-data-revolution_9789264234703-en. OECD (Organisation for Economic Co-operation and Partnership on AI. 2018. “Partners.” https://www.partnershiponai Development). 1980 (2013). “Guidelines on the Protection of .org/partners/. Privacy and Transborder Flows of Personal Data.” Paris. Pentland, Alex. 2015. Social Physics: How Social Networks Can http://www.oecd.org/sti/ieconomy/privacy.htm. Make Us Smarter. Reissue ed. New York: Penguin. ———. 2013. “The OECD Privacy Framework.” OECD, Paris. Peterson, Tim. 2015. “Facebook Users Are Posting 75% More http://www.oecd.org/internet/ieconomy/oecd_privacy Videos Than Last Year.” AdAge, January 7. http://adage.com _framework.pdf. / a r t i c l e / d i g i t a l / f a ce b o o k - u s e r s - p o s t i n g - 7 5 - v i d e o s ——— . 2015. “Data-Driven Innovation for Growth and Well- -year/296482/. Being: What Implications for Governments and Businesses.” Pfeifle, Sam. 2017. “Is the GDPR a Data Localization Law?” STI Policy Note, OECD, Paris. https://www.oecd.org/sti International Association of Privacy Professionals, /ieconomy/PolicyNote-DDI.pdf. Portsmouth, NH. https://iapp.org/news/a/is-the-gdpr-a-data ———. 2015. Digital Security Risk Management for Economic and -localization-law/. Social Prosperity: OECD Recommendation and Companion Piwek, Lukasz, David A. Ellis, Sally Andrews, and Adam Johnson. Document. Paris: OECD Publishing. https://doi.org/10.1787 2016. “The Rise of Consumer Health Wearables: Promises and /9789264245471-en. Barriers.” PLOS, February 2. http://journals.plos.org/plos ———. 2017. “Key Issues for Digital Transformation in the G20.” medicine/article?id=10.1371/journal.pmed.1001953. G20 Germany Presidency and OECD Conference, January 12. Poppe, Krijn J., Sjaak Wolfert, Cor Verdouw, and Tim Verwaart. https://www.oecd.org/g20/key-issues-for-digital-transforma 2013. Information and Communications Technologies as a tion-in-the-g20.pdf. Driver for Change in Agri-Food Chains. https://s3.amazonaws Ofcom. 2016. “Adults’ Media Use and Attitudes.” https://www .com/academia.edu.documents/40043142/542e8ba40cf27 7 . ofcom.org.uk/__data/assets/pdf_file/0026/80828/2016 d58e8ebffd.pdf20151115-68247-1ccjezq.pdf?AWSAcc ess -adults-media-use-and-attitudes.pdf. KeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1537343545 ———. 2018. “Ofcom Kicks Off Auction to Improve Mobile &Sig nature=ZMUDN6eXwMB17wLKmsIDXAZ Broadband and Prepare for 5G.” https://www.ofcom.org.uk zfR4%3D&re sponse-content-disposition=inline%3B%20 /about-ofcom/latest/media/media-releases/2018/start-spec filename%3DIn formation_and_Communication trum-auction. _Technology.pdf. Ohlhausen, Maureen K. 2017. “Remarks at the FTC Informational Price, R. 2017. “Mapping Apps Are Reportedly Directing People Injury Workshop.” December 12. https://www.ftc.gov/public Fleeing the Southern California Wildfires to Areas That Are -statements/2017/12/remarks-ftc-informational-injury on Fire.” Business Insider, December 7. http://www.businessin -workshop. sider.com/la-fires-gps-apps-directing-people-areas-fire-lapd -2017-12. Open Data Institute. n.d. “Data Infrastructure.” https://theodi.org /topic/data-infrastructure/. Priyadarshinee, Pragati, Rakesh Raut, Manoj Jha, and Bhaskar B. Gardas. 2017. “Understanding and Predicting the Open Government Working Group. n.d. “Open Government Determinants of Cloud Computing Adoption: A Two Staged Data.” http://opengovernmentdata.org/. Hybrid SEM-Neural Networks Approach.” Computers in Open Knowledge International. n.d. Open Data Handbook. http:// Human Behavior 76: 341–62. opendatahandbook.org. Qatar Ministry of Information and Communications Technology. ———. n.d. The Open Definition. http://opendefinition.org/. 2015. “Data Management Policy.” http://www.motc.gov.qa Packet Clearing house. n.d. “Internet Exchange Point Directory /sites/default/files/data_management_policy.pdf. Reports.” http://wwww.pch.net/ixpdir/summary. Ragoobar, Tricia, Jason Whalley, and David Harle. 2011. “Public Palmer, N. 2014. “Cracking Patterns in Big Data Saves Colombian and Private Intervention for Next-Generation Access Rice Farmers Huge Losses.” Research Program on Climate Deployment: Possibilities for Three European Countries.” Change, Agriculture, and Food Security (CCAFS), Montpellier, Telecommunications Policy 35 (9): 827–41. France. https://ccafs.cgiar.org/research/annual-report/2014 Rango, M., and M. Vespe. 2017. “Big Data and Alternative Data /cracking-patterns-in-big-data-saves-colombian-rice-farmers Sources on Migration: From Case Studies to Policy Support.” -huge-losses. Summary Report, European Commission. https://bluehub PARIS21 (Secretariat of the Partnership in Statistics for .jrc.ec.europa.eu/bigdata4migration/uploads/attachments Development in the 21st Century). 2015. A Road Map for /cjdelbdgo00hnqazv3u7xi6pd-big-data-workshop-draft-sum A Country-Led Data Revolution. Paris: OECD Publishing. mary-report.pdf. 142 Information and Communications for Development 2018 Reinsdorf, Marshall, and Gabriel Quiros. 2018. Measuring the Digital Sandvine. 2015. “Global Internet Phenomena: Latin America & Economy. Washington, DC: International Monetary Fund. North America.” https://www.sandvine.com/hubfs/down Reuters. 2017. “Rwanda Regulator Fines MTN Rwanda $8.5 Mln loads/archive/2015-global-internet-phenomena-report-latin over External IT Hub.” May 17. https://www.reuters.com -america-and-north-america.pdf. /article/rwanda-telecoms/rwanda-regulator-fines-mtn-rwan SAS Institute. n.d. “Big Data Analytics.” https://www.sas.com/en da-8-5-mln-over-external-it-hub-idUSL8N1IJ2IJ. _us/insights/analytics/big-data-analytics.html. Reuters Market Light. 2015. “Addressing Farmers Problems Schwab, Klaus. 2017. The Fourth Industrial Revolution. New York: through Digital India Initiative.” https://www.mygov.in Crown Business. / sites/ default/files/user_comments/Digital%20India Scott, Mark, and Mike Isaacs. 2016. “Facebook Restores Iconic -Agriculture.pdf. Vietnam War Photo It Censored for Nudity.” New York Times, Richter, Wolf. 2018. “Online Ad Revenues Are Surging, but 2 September 9. https://www.nytimes.com/2016/09/10/technol Companies Are Getting Most of the Spoils.” Business Insider, ogy/facebook-vietnam-war-photo-nudity.html?_r=0. April 27. http://www.businessinsider.com/online-ads-rev Seller, Marianne, and Paul Gray. 1999. “A Survey of Database enues-going-to-google-and-facebook-the-most-2017-4. Marketing.” Center for Research on Information Technology RICTA. 2015. “Growing the Rwandan Internet Content Hosted in and Organizations, University of California at Irvine. https:// Rwanda.” https://ricta.org.rw/sites/default/files/resources escholarship.org/uc/item/36z642kj. /1kwebsites_project_finalversion_0.pdf. Sengupta, Anirban, and Steven N. Wiggins. 2014. “Airline Pricing, Robertson, A. 2017. “What Makes YouTube’s Surreal Kids’ Videos Price Dispersion, and Ticket Characteristics On and Off the So Creepy?” The Verge, November 21. https://www.theverge Internet.” American Economic Journal: Economic Policy 6 (1): .com/culture/2017/11/21/16685874/kids-youtube-video 272–307. -elsagate-creepiness-psychology. Sengupta, Ritam, Richard Heeks, Chattapadhyay Sumandro, and Rochet, Jean-Charles, and Jean Tirole. 2003. “Platform Competition Christopher Foster. 2017. “Exploring Big Data for in Two-Sided Markets.” European Economic Association 1 (4): Development: An Electricity Sector Case Study from India.” 990–1029. Working paper, University of Manchester, Manchester, UK. Seybert, Heidi, and Petronela Reineckle. n.d. “Internet and Cloud Rohlfs, J. 1974. “Theory of Interdependent Demand for a Services: Statistics on Use by Individuals.” EuroStat. http:// Communications Service.” Bell Journal of Economics and ec.europa.eu/eurostat/statistics-explained/index.php Management Science 5 (1): 16–37. /Internet_and_cloud_services_-_statistics_on_the_use_by Romano, Aja. 2018. “How Facebook Made It Impossible to Delete _individuals. Facebook.” VOX, March 22. https://www.vox.com/culture Shehabi, A., S. J. Smith, N. Horner, I. Azevedo, R. Brown, /2018/3/22/17146776/delete-facebook-difficult. J. Koomey, E. Masanet, D. Sartor, M. Herrlin, and W. Lintner. Roosendaal, Arnold, Marc van Lieshout, and Anne Fleur van 2016. “United States Data Center Energy Usage Report.” Veenstra. 2014. “Personal Data Markets.” TNO Innovation for Lawrence Berkeley National Laboratory, Berkeley, CA. https:// Life. http://publications.tno.nl/publication/34612412/riZsP9 pubarchive.lbl.gov/islandora/object/ir%3A1005775 /TNO-2014-R11390.pdf. /datastream/PDF/view. Rosenwald, Michael. 2015. “The Digital Media Industry Needs to Sherman, Alex. 2018. “Uber Is Preparing to Sell Southeast Asia React to Ad Blockers . . . or Else.” Colombia Journalism Review, Unit to Grab in Exchange for Stake in Company: Sources.” September/October. https://www.cjr.org/business_of_news CNBC News, February 16. https://www.cnbc.com /will_ad_blockers_kill_the_digital_media_industry.php. /2018/02/16 /uber-preparing-to-sell-southeast-asia-unit-to Russell, Stuart, and University of California at Berkeley. 2017. -grab.html. “Statement at the ITU AI for Global Good Summit.” Geneva, Shu, Catherine. 2018. “Waze Signs Data-Sharing Deal with June. https://www.youtube.com/watch?v=FsSfB8aKTuI. AI-Based Traffic Management Startup Waycare.” Tech Crunch, Ruutu, Sampsa, Thomas Casey, and Ville Kotivirta. 2017. April 26. https://techcrunch.com/2018/04/26/waze-signs “Development and Competition of Digital Service Platforms: -data-sharing-deal-with-ai-based-traffic-management-star A System Dynamics Approach.” Technological Forecasting and tup-waycare/. Social Change 117: 119–30. Simonite, Tim. 2014. “Sell Your Personal Data for $8 a Month.” Salem, Fadi. 2017. The Arab World Online: Digital Transformations MIT Technology Review, February 12. https://www.technol and Societal Trends in the Age of the 4th Industrial Revolution. ogyreview.com/s/524621/sell-your-personal-data-for-8-a Vol. 3. Dubai: Mohammed Bin Rashid School of Government. -month/. https://img0bm.b8cdn.com/images/uploads/article_docs Simons, Gary F., and Charles D. Fenning, eds. 2018. “English.” /arab_world_online_2017_4ir_d_fadi_salem_34792_EN.pdf. Ethnologue: Languages of the World. 21st ed. English. Bibliography 143 Dallas, TX: SIL International. https://www.ethnologue.com Multi-sided Platforms: The Digital Ecosystem Strategy of /language/eng. Alibaba.com.” Journal of the Association for Information Singer, Natasha. 2012. “Mapping, and Sharing, the Consumer Systems 16 (4): 248–80. Genome.” New York Times, June 17. https://www.nytimes Täuscher, Karl. n.d. “Business Models in the Digital Economy: An . com/2012/06/17/technology/acxiom-the-quiet-giant-of Empirical Study of Digital Markets.” Working paper, -consumer-database-marketing.html. Fraunhofer MOEZ, Fraunhofer Center for International Sloane, Garett. 2018. “What Facebook Crisis? Ad Sales Skyrocket to Management and Knowledge Economy, Leipzig. https://www $11.8 Billion, Users Don’t #delete.” AdAge, April 25. http:// .imw.fraunhofer.de/content/dam/moez/de/documents adage.com/article/digital/privacy-concerns-facebook-ad / Working_Paper/Working_Paper_Digital_Marketplaces -sales-hits-11-8-billion/3s13271/. _final.pdf. Smith, Eve. 2018. “A Memo to Big Tech: The Techlash against Täuscher, Karl, and Sven M. Laudien. 2017. “Uncovering the Amazon, Facebook and Google—And What They Can Do.” Nature of Platform-Based Business Models: An Empirical The Economist, January 20. https://www.economist.com Taxonomy.” Paper presented at the 50th Hawaii International /briefing/2018/01/20/the-techlash-against-amazon-facebook Conference on System Sciences (HICSS), Manoa, HI. https:// -and-google-and-what-they-can-do. scholarspace.manoa.hawaii.edu/bitstream/10125/41802/1 /paper0653.pdf. Statistics New Zealand. 2016. “Statistics New Zealand’s Strategic Intentions for the Period 2016/17–19/20 Annual Report for ———. 2018. “Understanding Platform Business Models: A the Year Ended 30 June 2016.” Wellington. https://www.stats Mixed Methods Study of Marketplaces.” European Management .govt.nz/about-us/corporate-publications/. Journal 36 (3): 319–29. Still, Kaisa, Marko Seppänen, Heidi Korhonen, Katri Valkokari, TechMoran. 2017. “Jumia Launches Offline Experiential Centers Arho Suominen, and Miika Kumpulainen. 2017. “Business in Kenya.” September 14. https://techmoran.com/jumia Model Innovation of Startups Developing Multisided Digital -launches-offline-experiential-centers-in-kenya/. Platforms.” In 2017 IEEE 19th Conference on Business Telecom Regulatory Authority of India. 2016. “TRAI Releases the Informatics (CBI 2017), Vol. 2, 70–75. http://ieeexplore.ieee Prohibition of Discriminatory Tariffs for Data Services . org/stamp/stamp.jsp?tp=&arnumber =8012942&isnum Regulations.” http://www.trai.gov.in/notifications/press ber=8012382. -release/trai-releases-prohibition-discriminatory-tariffs-data Stuart, Elizabeth, Emma Samman, William Avis, and Tom Berliner. -services. 2015. The Data Revolution: Finding the Missing Millions. Thailand National Statistical Office. 2016. “The 2016 Household Research Report Series, Overseas Development Institute. Survey on the Use of Information and Communication London: Development Progress. https://www.odi.org/sites/odi Technology.” Ministry of Information and Communication .org.uk/files/odi-assets/publications-opinion-files/9604.pdf. Technology. http://web.nso.go.th/en/survey/ict/ict_house16 Suominen, Kati. 2017. “Accelerating Digital Trade in Latin America .htm. and the Caribbean.” Working paper, Inter-American The Economist. 2017. “The World’s Most Valuable Resource Is No Development Bank, Washington, DC. https://publications Longer Oil, but Data.” May 6. https://www.economist.com .iadb.org/bitstream/handle/11319/8166/Accelerating-Digital / news/leaders/21721656-data-economy-demands-new -Trade-in-Latin-America-and-the-Caribbean.PDF. -approach-antitrust-rules-worlds-most-valuable-resource. Surowiecki, James. 2016. “Why Startups Are Struggling.” Technology ———. 2018. “A Bold Scheme to Dominate Ride-Hailing.” May Review, June 15. https://www.technologyreview.com/s/601497 10. https://www.economist.com/briefing/2018/05/10/a-bold /why-startups-are-struggling/. -scheme-to-dominate-ride-hailing. Swanson, Alexandra, Margaret Kosmala, Chris Lintott, Robert ———. 2018. “How to Fix What Has Gone Wrong with the Simpson, Arfon Smith, and Craig Packer. 2015. “Snapshot Internet.” Special Report, June 28. https://www.economist Serengeti, High-Frequency Annotated Camera Trap Images .com/special-report/2018/06/28/how-to-fix-what-has-gone of 40 Mammalian Species in an African Savanna.” -wrong-with-the-internet. Nature. Scientific Data. https://www.nature.com/articles/ ———. 2018. “The Techlash against Amazon, Facebook and sdata201526. Google—And What They Can Do.” January 20. https://www Syntonic. 2017. “Syntonic Expands to Latin America and Lending .economist.com/briefing/2018/01/20/the-techlash -against Mobile Carriers.” January 21. https://syntonic.com/2017/01 -amazon-facebook-and-google-and-what-they-can-do. / syntonic-expands-to-latin-america-with-leading-mobile The Green Grid. 2016. “The ‘Big Bad Data Police’ Will Call If -carriers/. We Don’t Get More Efficient.” News release, September 5. Tan, Barney, Shan Ling Pan, Ziangua Lu, and Lihua Huang. https://www.thegreengrid.org/en/newsroom/news-releases 2015. “The Role of IS Capabilities in the Development of /green-grid-big-bad-data-police-will-call. 144 Information and Communications for Development 2018 The Guardian. n.d. “The Cambridge Analytica Files.” https://www ———. 2017. “Social Media and Forced Displacement: Big Data .theguardian.com/news/series/cambridge-analytica-files. Analytics and Machine-Learning.” White paper. http://unglob Thornhill, John. 2018. “Social Media Users of the World Unite.” alpulse.org/sites/default/files/White%20Paper%20Social%20 Financial Times, February 5. https://www.ft.com/content Media%203_0.pdf. /ea6c3a0c-0843-11e8-9650-9c0ad2d7c5b5. ———. 2017. “Using Machine Learning to Analyse Radio Content Toonders, J. 2014. “Data Is the New Oil of the Digital Economy.” in Uganda.” http://unglobalpulse.org/sites/default/files/Radio Wired. https://www.wired.com/insights/2014/07/data-new %20Analysis%20Report_Preview%20%283%29.pdf. -oil-digital-economy/. ———. 2018. “Data Innovation Risk Assessment Tool.” Data Touger, Glenn Evan. 2018. “What’s the Difference between Innovation for Development Guide. http://unglobalpulse.org Artificial Learning (AI), Machine Learning, and Deep /sites/default/files/Privacy%20Assessment%20Tool%20.pdf. Learning?” Prowesscorp, August 3. www.prowesscorp.com UNCTAD (United Nations Conference on Trade and /whats-the-difference-between-artificial-intelligence-ai-mach Development). 2015. “Review of E-Commerce Legislation ine-learning-and-deep-learning/. Harmonization in the Economic Community of West African Tsai, Wen Chun. 2016. “Analyzing the Emergence of Alibaba Group States.” UNCTAD, Geneva and New York. http://unctad.org from Business Ecosystem Perspective.” Journal of International /en/PublicationsLibrary/dtlstict2015d2_en.pdf. Management Studies 11 (2): 53–64. ———. 2016. “Data Protection Regulations and International Tucker, C. 2017. “Privacy, Algorithms and Artificial Intelligence.” Data Flows: Implications for Trade and Development.” Working Paper 14011, National Bureau of Economic UNCTAD, Geneva and New York. http://unctad.org/en Research, Cambridge, MA. http://www.nber.org/chapters /PublicationsLibrary/dtlstict2016d1_en.pdf. /c14011.pdf. ———. 2017. Information Economy Report 2017: Digitalization, Turovsky, Barak. 2016. “Found in Translation: More Accurate, Trade and Development. Geneva: UNCTAD. http://unctad Fluent Sentences in Google Translate.” Google, November 15. .org/en/PublicationsLibrary/ier2017_en.pdf. https://blog.google/products/translate/found-translation ———. 2018. “Data Protection and Privacy Legislation -more-accurate-fluent-sentences-google-translate/. Worldwide.” UNCTAD, Geneva and New York. http://unctad UN Broadband Commission. 2017. “Working Group on Education: .org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom Digital Skills for Life and Work.” Report, Broadband -Data-Protection-Laws.aspx. Commission for Sustainable Development. http://unesdoc UNDG (United Nations Development Group). 2017. “Data .unesco.org/images/0025/002590/259013e.pdf. Privacy, Ethics, and Protection: Guidance Note on Big Data UN Data Revolution Group. 2014. “A World That Counts: and the Achievement of the 2030 Agenda.” https://undg.org Mobilising the Data Revolution for Sustainable Development.” /wp-content/uploads/2017/11/UNDG_BigData_final_web http://www.undatarevolution.org/wp-content/uploads .pdf. /2014/12/A-World-That-Counts2.pdf. ———. 2017. “Guidance Note on Big Data for Achievement of the UN Global Pulse. 2014. “Nowcasting Food Prices in Indonesia 2030 Agenda: Data Privacy, Ethics and Protection.” https:// Using Social Media Signals.” Global Pulse Project Series, no. 1. undg.org/wp-content/uploads/2017/11/UNDG_BigData https://www.unglobalpulse.org/projects/nowcasting-food _final_web.pdf -prices. UNDP (United Nations Development Programme) and UN ———. 2015. Using Twitter Data to Analyze Public Sentiment on Global Pulse. 2016. A Guide to Data Innovation for Fuel Subsidy Policy Reform in El Salvador. Global Pulse Project Development: From Idea to Proof of Concept. http://www.undp Series, no. 13. www.unglobalpulse.org/sites/default/files/UNGP .org/content/undp/en/home/librarypage/development-im _ProjectSeries_ElSalvador_Fuel_2015_0.pdf. pact/a-guide-to-data-innovation-for-development---from ———. 2016. “A Guide to Data Innovation for Development: -idea-to-proof-.html. From Idea to Proof of Concept.” http://unglobalpulse.org UNESCO (United Nations Educational, Scientific, and Cultural /sites/default/files/UNGP_BigDataGuide2016_%20Web.pdf. Organization). 2018. A Lifeline to Learning: Leveraging ———. 2016. “Haze Gazer: A Crisis Analysis Tool.” Tool Series, Technology to Support Education for Refugees. Paris: UNESCO. no. 2. http://unglobalpulse.org/sites/default/files/2-pager%20 http://unesdoc.unesco.org/images/0026/002612/261278e.pdf. Haze%20Gazer%20-%20Feb%202017_0.pdf. UNICEF. 2015. “Procedure for Ethical Standards in Research, ———. 2017. “Inferring Commuting Statistics in the Greater Evaluation, Data Collection and Analysis.” CF/PD/DRP/2015- Jakarta Area with Twitter.” https://www.unglobalpulse.org 001. April 1. https://www.unicef.org/supply/files / projects/inferring-commuting-statistics-greater-jakarta / ATTACHMENT_IV-UNICEF_Procedure_for_Ethical -area-twitter. _Standards.PDF. Bibliography 145 UNICEF and Facebook. n.d. “A Case Study: Data and Social Media Virtu. 2015. “5 Data Security Challenges Faced by Government Can Lead to Healthier Lives.” Infographic. http://neo-assets Agencies, and What They Can Do about It.” September 23. .s3.amazonaws.com/news/FB-UNICEF-Big.png. https://www.virtru.com/blog/data-security/. United Nations. 2015. Transforming Our World: The 2030 Agenda Wang, Teresa, and Malay Gandhi. 2014. “The Future of Biosensing for Sustainable Development. New York: UN Publishing. Wearables: A Review of the Current Landscape and Future of https://sustainabledevelopment.un.org/post2015/transfor Biosensing Wearables.” Rock Health. https://rockhealth.com mingourworld. /reports/the-future-of-biosensing-wearables/. Uptime Institute. 2012. “Data Center Site Infrastructure Tier Wang, W., and F. Guo. 2016. “RoadLab: Revamping Road Condition Standard: Topology.” http://www.gpxglobal.net/wp-content and Road Safety Monitoring by Crowdsourcing with /uploads/2012/10/TIERSTANDARD_Topology_120801.pdf. Smartphone App.” Paper presented at the 95th Annual Meeting of the Transportation Research Board, Washington, Upwork. 2018. “Upwork Releases Q4 2017 Skills Index, Ranking DC, January 10–14. the 20 Fastest-Growing Skills for Freelancers.” Press release, February 7. https://www.upwork.com/press/2018/02/07/q4 Warren, Mark. 2016. “The Cure for Cancer Is Data—Mountains of -2017-skills-index/. Data.” Wired, October 19. https://www.wired.com/2016/10 /eric-schadt-biodata-genomics-medical-research/. United States, Executive Office of the President. 2016. Big Data: A Report on Algorithmic Systems, Opportunity, and Civil Rights. Warren, Samuel D., and Louis D. Brandeis. 1890. “The Right to Washington, DC: White House. https://www.whitehouse.gov Privacy.” Harvard Law Review 4 (5): 193. / sites/default/files/microsites/ostp/2016_0504_data Wesolowski, A., T. Qureshi, M. F. Boni, P. R. Sundsøy, M. A. _discrimination.pdf. Johansson, S. B. Rasheed, K. Engø-Monsen, and C. O. Buckee. United States, White House, Office of the Press Secretary. 2016. 2015. “Impact of Human Mobility on the Emergence of “Fact Sheet: Data by the People, for the People—Eight Years of Dengue Epidemics in Pakistan.” National Academy of Sciences Progress Opening Government Data to Spur Innovation, 112 (38): 11887–92. Opportunity, and Economic Growth.” White House, WFP (World Food Programme). 2015. WFP Guide to Personal Washington, DC. September 28. https://www.whitehouse.gov Data Protection and Privacy. Rome: WFP. https://docs.wfp /the-press-office/2016/09/28/fact-sheet-data-people-people .org/api/documents/e8d24e70cc11448383495caca154cb97 -eight-years-progress-opening-government. /download/. USA Today. 2017. “Rebel Wilson Wins Defamation Case against Whitler, Kimberly A. 2016. “The Personal Data Revolution: Why Publisher.” June 15. https://www.usatoday.com/story/life It’s Time for Marketers to Care.” Forbes, September 18. https:// / people/2017/06/15/rebel-wilson-wins-defamation-case www.forbes.com/sites/kimberlywhitler/2016/09/18/the-per -against-publisher/102877120/. sonal-data-revolution-why-its-time-for-marketers-to -care/#488b46db5681. Van Alstyne, M. 2016. “Platform Shift: How New Biz Models Are Changing the Shape of Industry.” Video. https://www Wikipedia. n.d. “DARPA Network Challenge.” https://en.wikipedia .youtube.com/watch?v=8OFRD66pI0Y. .org/wiki/DARPA_Network_Challenge. van Zyl, Gareth. 2014. “MTN Unveils Africa’s First Solar Data Williams, Chris, Davide Strusani, David Vincent, and David Kovo. Centre Cooling System.” ITWeb Africa, July 10. http://www 2016. “The Economic Impact of Next-Generation Mobile . itwebafrica.com/cloud/517-south-africa/233194-mtn-un Services: How 3G Connections and the Use of Mobile Data veils-africas-first-solar-data-centre-cooling-system. Impact GDP Growth.” Chapter 1.6 in The Global Information Vector. 2017. “Operational Performance for the 9 Months Ended Technology Report. Geneva: World Economic Forum. http:// 31 March 2017.” Market release, April 26. https://www.vector www3.weforum.org/docs/GITR/2013/GITR_Chapter1.6 .co.nz/news/operational-performance-for-the-9-months _2013.pdf. -ended-31. Wong, Julia Carrie. 2016. “Mark Zuckerberg Accused of Abusing Verge, Jason. 2015. “PLDT Building Eighth Philippines Data Power after Facebook Deletes ‘Napalm Girl’ Post.” The Center for $29m.” Data Center Knowledge, April 28. http:// Guardian, September 9. https://www.theguardian.com/tech www.datacenterknowledge.com/archives/2015/04/28/pldt nology/2016/sep/08/facebook-mark-zuckerberg-napalm-girl -building-eighth-philippines-data-center-29m. -photo-vietnam-war. Verhulst, Stefaan. 2014. “Mapping the Next Frontier of Open Data: World Bank. 2008. “P106589: Mexico Information Technology Corporate Data Sharing.” Blog, United Nations Global Pulse, (IT) Industry Development Project.” World Bank, Washington, September 17. http://unglobalpulse.org/mapping-corporate DC. http://projects.world bank.org/P106589/information -data-sharing. -technology-development?lang=en. 146 Information and Communications for Development 2018 ———. 2014. “Open Data for Economic Growth.” Transport and ———. 2017. “RAM—Rural Accessibility Hub. GitiHub.” ICT Global Practice, Washington, DC. World Bank, World Bank, Washington, DC. https://github.com/WorldBank Washington, DC. http://www.worldbank.org/content/dam -Transport/ram-backend. /Worldbank/document/Open-Data-for-Economic-Growth ———. 2018. Digital Adoption Index. 2018 update. http://www.world .pdf. bank.org/en/publication/wdr2016/Digital-Adoption-Index. ———. 2016. “Philippines: Real-Time Data Can Improve Traffic ———. n.d. India Lights Platform. http://india.nightlights.io/. Management in Major Cities.” Press Release, April 5. World Bank, Washington, DC. http://www.worldbank.org/en/news World Bank and CGAP (Consultative Group to Assist the Poor). / press-release/2016/04/05/philippines-real-time-data 2018. “Data Protection and Privacy for Alternative Data.” -can-improve-traffic-management-in. Draft GPFI-FCPL Sub-Group Discussion Paper. Group of Twenty (G20). https://www.g20.org/sites/default/files/docu ———. 2016. Big Data: Innovation Challenge: Pioneering mentos_producidos/data_protection_and_privacy_for Approaches to Data-Driven Development. Washington, DC: _alternative_data_wbg_0.pdf. World Bank. World Bank Group. 2018. Africa’s Pulse, no. 17 (April). https:// ———. 2016. World Development Report 2016: Digital Dividends. www.openknowledge.worldbank.org/handle/10986/29667. Washington, DC: World Bank. doi:10.1596/978-1-4648-0671-1. World Economic Forum. 2011. “Personal Data: The Emergence of ———. 2017. Big Data in Action for Government: Big Data a New Asset Class.” World Economic Forum, Geneva. http:// Innovation in Public Services, Policy, and Engagement. www3.weforum.org/docs/WEF_ITTC_PersonalData New Washington, DC: World Bank. http://documents.worldbank Asset_Report_2011.pdf. . org/curated/en/176511491287380986/Big-data-in-action -for-government-big-data-innovation-in-public-services ———. 2013. “Unlocking the Value of Personal Data: From -policy-and-engagement. Collection to Usage.” World Economic Forum, Geneva. http:// www3.weforum.org/docs/WEF_IT_UnlockingValuePersonal ———. 2017. Doing Business 2017: Equal Opportunity for All. Data_CollectionUsage_Report_2013.pdf. Washington, DC: World Bank. http://www.doingbusiness.org /reports/global-reports/doing-business-2017. Wright, Morgan. 2018. “A Ransomware Attack Brought Atlanta to ———. 2017. “Internet of Things: The New Government to Its Knees—And No One Seems to Care.” The Hill, April 4. Business Platform—A Review of Opportunities, Practices, http://thehill.com/opinion/cybersecurity/381594-a-ransom and Challenges.” World Bank, Washington, DC. http://docu ware-attack-brought-atlanta-to-its-knees-and-no-one ments.worldbank.org/curated / en/610081509689089303 -seems-to. /Internet-of-things-the-new -government-to-business-plat Zannier, Federico. n.d. “A Bite of Me.” https://www.kickstarter form-a-review-of-opportunities-practices-and-challenges. .com/projects/1461902402/a-bit-e-of-me/description. Bibliography 147 Contributors Mona Badran (World Bank, Digital Development, the Institute of Chartered Accountants in England and consultant)—chapter 4 Wales, and a master’s in Economics for Development Associate Professor, Faculty of Economics and Political from Oxford University. Prior to joining the ITU, Phil- Science, Cairo University, Egypt lippa worked as an Economic Affairs Officer at the United Nations Conference on Trade and Development and as a Mona is an economist with a main focus on digital econom- consultant with the United Nations Industrial Development ics, economics of telecommunications, and information and Organization in Tanzania and Egypt. She is chief author of communication technology’s role in development. She has the State of Broadband report, Confronting the Crisis reports extensive consulting experience working for numerous private and Fast-Forward Progress report. She analyzes develop- sector clients and international organizations, such as the ments in broadband, voice over internet protocol, and 3G World Bank, International Telecommunication Union (ITU), markets around the world. GSMA, United Nations Economic and Social Commission for West Asia, and International Labour Organization. For Eva Clemente Miranda (World Bank, Finance more than five years, she advised the Egyptian Ministry of Competitiveness and Innovation)—chapters 5 and 6 Investment, where she led the research department. Numerous Eva, a Spanish national, started her journey at the World times, she was awarded Cairo University’s International Publi- Bank Group in early 2012 when she joined the Digital cations Award for publishing in Thomson Reuters–indexed Development Unit. As an Information and Communica- journals in the area of Information and Communication Tech- tion Technology (ICT) Policy Specialist, she has assisted nologies for Development. client countries in their efforts to harness the benefits of ICT-enabled innovation. As data has become an essential Phillippa Biggs (International Telecommunication ingredient for innovation in today’s digital economy, Eva’s Union)—chapter 3 work has increasingly focused on supporting data policies Senior Policy Analyst, ITU and data-driven innovation and entrepreneurship in Latin Phillippa has been an economist and qualified accountant America and Africa at the national and subnational levels. with the International Telecommunication Union (ITU) Currently, in the Finance, Competitiveness, and Inno- since 2005. She holds a Natural Sciences degree from vation Global Practice, she promotes the digital economy Cambridge University, an accountancy qualification from agenda for Africa. 149 Elena Gasol Ramos (World Bank, Finance Competi- Christine Howard (World Bank, Digital tiveness and Innovation)—chapters 5 and 6 Development)—Bibliography and Publication Elena, a lawyer by training, is a Senior Private Sector Christine is a Program Assistant with the Digital Develop- Specialist with the World Bank, based in Washington, ment Department at the World Bank. She joined the World DC. She currently leads the Bank’s engagement on digital Bank staff in 2012 after graduating with her bachelor’s economy, entrepreneurship, and innovation in Kenya and in Political Science and Creative Writing earlier that year. Burundi. She also provides advice to governments on data Since joining the staff, she has supported and contributed protection issues. Prior to this her work included a variety of to multiple digital development–related research, programs, positions within and outside the World Bank Group, in both and initiatives. She creates original art, short stories, and Europe and the United States. Her areas of expertise include poems. Her poetry has been published in The Dulcimer, a data protection and data security, consumer protection, student-led literary and art magazine. and ICT policy and regulation. She is a member of both the Salamanca and New York bars and has taught comparative Tim Kelly (World Bank, Digital Development)—Editor, privacy law, with a focus on e-privacy, at the Georgetown Executive Summary, and chapters 1, 2, and 5 University Law Center. She has a master’s in European Law Tim is a Lead Digital Development Specialist based in from the College of Europe in Bruges, Belgium, and an LLM Nairobi. He is the editor of this report and the over- from Georgetown University Law Center. all Information and Communications Technology for Development series. He worked previously at the Organi- Rachel Firestone (Hala Systems—formerly World sation for Economic Co-operation and Development and Bank, Digital Development)—chapter 4 International Telecommunication Union, having joined Rachel is the Director of Operations at Hala Systems, a social the World Bank staff in 2008. His other World Bank enterprise that uses advanced technologies to save lives publications include the policy chapter in the 2016 World in conflict zones, to combat disinformation, and to bring Development Report, Maximizing Mobile, and ICTs for accountability for war crimes. She specializes in informa- Post-conflict Reconstruction, as well as the Broadband Strat- tion and communication technology (ICT) and sustainable egies Handbook (with Carlo Rossotto). In addition to his development for communities in conflict and spent several analytical work and technical assistance, he is also co–task years with the World Bank working on projects in Somalia team leader for digital development investment lending and the Horn of Africa. Prior to joining the World Bank programs in Comoros, Ghana, Malawi, Niger, Tanzania, staff, she spent four years in India working on self-advocacy and Somalia. and protection programs with communities affected by conflict and natural disaster. Rachel completed her master’s at Georgetown University, with a concentration in Global Robert Kirkpatrick (UN Global Pulse)—chapter 3 Politics and Security. Robert is the Director of UN Global Pulse, a UN initiative driving a big data revolution for global development and Roku Fukui (World Bank, Digital Development, resilience. Prior to joining the UN staff, Robert cofounded consultant)—Executive Summary and chapter 4 and led software development for two pioneering private Roku works as a consultant with the World Bank Group’s sector humanitarian technology teams, first at Groove Digital Development Department. He focuses on various Networks, and later as Lead Architect for Microsoft aspects of digital development and has worked primar- Humanitarian Systems. From 2007 to 2009 he served as ily in Afghanistan and Somalia. His research interests Chief Technology Officer of the nonprofit InSTEDD. cover information and communication technology for Robert was a member of the UN Secretary-General’s development and mobile innovation. Roku received his Independent Expert Advisory Group on a Data Revolution MA in International Economics and International Rela- for Sustainable Development (2014) and currently sits tions from the Johns Hopkins School of Advanced Inter- on InSTEDD’s Board of Directors, as well as the World national Studies. He is an above-average cryptocurrency Economic Forum’s Global Agenda Council on Data- investor. Driven Development. 150 Information and Communications for Development 2018 Prasanna Lal Das (World Bank, Finance, Competi- Michael Minges (World Bank, Digital Development, tiveness, and Innovation)—chapters 5 and 6 consultant)—chapters 2, 5, and 6 and Data Notes Prasanna works on entrepreneurship, data/digital strategy, Michael Minges is the lead consultant at ICTData, where he and disruptive technologies in the Finance, Competitiveness provides advice and analysis on digital technology issues for and Innovation Global Practice at the World Bank Group. a range of clients including governments, the private sector, His most recent publication is “Internet of Things—The and international organizations. He previously worked at the Next Government to Business Platform.” His current work International Telecommunication Union and International includes projects on technologies such as blockchain, Inter- Monetary Fund. Michael drafted the technology chapter for the net of Things, and machine learning applied to development World Bank’s Broadband Strategies Toolkit. Recent assignments questions such as financial inclusion, growth of small and include analyzing the impact of broadband in least developed medium-sized enterprises, and entrepreneurship ecosystem countries, developing an e-commerce strategy for Oman, and diagnostics. Prasanna managed the Bank Group’s open evaluating the Taza Koom digital transformation initiative in financial data program and led the development of its open the Kyrgyz Republic. He holds an MBA from George Washing- trade and competitiveness data platform. Prasanna holds ton University. a master’s degree in Modern Indian History. He can be followed on Twitter @prasannalaldas. Tatiana Nadyseva (World Bank, Digital Development, consultant)—chapter 4 Bradley Larson (World Bank, East Asia Pacific, Tatiana joined the World Bank as a consultant to work on consultant)—Data Notes the “People and Data” chapter for this report. She previously Bradley Larson is a consultant with the Macroeconomics, worked in the sphere of advocacy, gender equality, and infor- Trade, and Investment Global Practice, working primar- mation and communication technology–enabled employ- ily on issues related to the digital economy of East Asia. ment, but her real passion lies in the sphere of technological Previously, he led data analysis and visualization for three evolution. She received her BSc and postgraduate diploma in World Development Reports: Learning to Realize Educa- Economics from Saint Petersburg University of Engineering tion’s Promise (2018), Governance and the Law (2017), and and Economics and her MSc in Technology (Operations and Digital Dividends (2016). He has also worked for the World Innovation Management) from Aalborg University. Apart Bank’s Public Sector Governance unit, the Special Inspector from being a tech enthusiast, Tatiana is also an environment- General for Iraq Reconstruction, and the Center for Strategic alist, vegan, digital nomad, and photographer. and International Studies. He has an MA in International Economics and Strategic Studies from the Johns Hopkins Siddhartha Raja (World Bank, Digital School of Advanced International Studies. Development)—chapter 4 Siddhartha Raja is a Digital Development Specialist with Miguel Luengo-Oroz (UN Global Pulse)—chapter 3 the World Bank Group. His work focuses on connecting Miguel is the Chief Data Scientist at UN Global Pulse. As the more people to better and cheaper internet and digital first data scientist at the United Nations, since 2011, Miguel technologies. He has assisted governments in designing has created and managed teams that have implemented and implementing policy reform and investment programs more than 30 innovation projects worldwide with govern- that have expanded broadband connectivity, helped people ments and UN agencies. He also advises the government of develop their digital skills and find work online, and gener- Spain in regard to its artificial intelligence strategy. He is the ated exponential improvements in international connec- founder of MalariaSpot.org at the Universidad Politecnica tivity, bringing people closer to information, markets, and de Madrid—video games and crowdsourcing for medical public services. He has a bachelor’s degree in Telecom- diagnosis. Over the last 15 years, he has coauthored more munications Engineering from the University of Bombay than 40 scientific publications. Prior to joining the United and a master’s degree in Infrastructure Policy Studies from Nations, Miguel worked as an antidisciplinary scientist in Stanford University, has studied media law and policy at the French and Spanish institutions in fields like artificial crea- University of Oxford, and has a doctorate in Telecommuni- tivity and genetics. cations Policy from the University of Illinois. Contributors 151 Liudmyla (Mila) Romanoff (UN Global Pulse)— on broadband, technology, and development in over 40 chapter 3 engagements, including the Russian Federation, Ukraine, Mila is the Legal and Privacy Specialist at the UN Global the European Union, the Arab Republic of Egypt, Algeria, Pulse, where she leads the data privacy and risk manage- Tunisia, Morocco, and Jordan and in West Bank and Gaza, ment program and is responsible for establishing legal Cambodia, Bosnia and Herzegovina, Libya, and Iraq. He mechanisms for public-private data partnerships. Mila is worked previously at the Inter-American Development the lead drafter of the first UN system–wide framework on Bank and in management consulting, advising leading data privacy and digital ethics, formally adopted by United European technology firms on demand analysis, market- Nations Development Group. She currently coordinates the ing, corporate strategy, and regulatory affairs. He holds UN Global Pulse Data Privacy Advisory Group and cofacili- postgraduate degrees in Economics and Business Admin- tates the UN Privacy Policy Group. Previously, she advised istration from Bocconi University in Milan and in Finan- two permanent country missions to the UN and worked in cial and Commercial Regulation from the London School the private sector as a commercial contracting and litigation of Economics. attorney. Mila sits on several legal and privacy associations and privacy advisory boards. She is licensed to practice law Felicia Vacarelu (UN Global Pulse)—chapter 3 in Ukraine and New York. Felicia leads communications and social media activities for UN Global Pulse, manages media outreach, and helps build Carlo Rossotto (World Bank, Digital Development)— and maintain fruitful relationships with partners. Over the chapter 5 past eight years, she has worked with various UN offices Carlo Maria Rossotto is a Lead ICT Specialist at the and departments, coming to Global Pulse from the Food World Bank and leads the Digital Economy Window of and Agriculture Organization in Rome. Prior to working the Digital Development Partnership, the Bank’s new for the United Nations, Felicia was Media and Outreach Trust Fund Facility bringing together governments and Coordinator for the 2009 Black Sea Energy and Economic leading technology firms to foster digital development. Forum in Romania. She also held several editorial and At the World Bank, he has been responsible for lend- journalistic positions at Mediafax, one of Romania’s most ing and technical assistance operations in Europe, the prestigious news agencies. To strengthen her knowledge Middle East, North Africa, and East Asia. Carlo is one in the field, she is currently pursuing a master’s in Media of the Bank’s leading authorities on broadband and the and Public Relations from the University of Leicester in the digital economy. He has advised top-level policy makers United Kingdom. 152 Information and Communications for Development 2018