Cloud Empires: How Digital Platforms Are Overtaking the State and How We Can Regain Control

1

INTRODUCTION

An American company was refusing to pay a Bangladeshi supplier, arguing that the delivery wasn’t matching their expectations. The amount of money at stake was trivial by US standards but could feed a family for a week in Bangladesh. The supplier had brought an action against the client. It was now Zara Khan’s job to examine the evidence, hear the parties’ arguments, and adjudicate.¹

But Khan was not a judge: she was a dispute resolution agent working for a major digital marketplace platform. She didn’t have any legal training as such. She came from a pretty modest background and had worked as a virtual assistant before landing her current gig. The technology company had trained her on the laws of its marketplace and taught her how to apply them to different types of cases. Now she was resolving disputes worth many thousands of dollars every day.

Most cases were easily resolved. Often defendants didn’t respond to a complaint at all within the allotted thirty-day period, perhaps knowing that their case was weak. When that happened, the system automatically issued a default decision in favor of the complainant and released funds from an escrow account where they sat pending the conclusion of the transaction. Khan then moved on to the next complaint.

In this case, though, the defendant was arguing his case vigorously—in ALL CAPS. Khan set aside what she thought about that style of argument and examined the facts of the case in light of the marketplace’s rules. She was expected to handle all cases meticulously and impartially so as to safeguard users’ trust in the platform. That said, she was also expected to process up to forty cases a week to max out her performance bonuses.

The process was that she would first try to get the parties to settle the case amicably. She’d mediate between them, explaining at each turn what the applicable rules were, hoping that they’d begin to see eye to eye. In this case, though, the parties’ emotions were running high. An amicable settlement seemed unlikely.

Once thirty days had passed without settlement, it was time for Khan to issue a decision. If she ruled in favor of the Bangladeshi supplier, then the funds would belong to the supplier. If she ruled in favor of the US client and there was evidence of deception on the part of the supplier, then she had the power to throw out the supplier from the marketplace permanently.

But in certain types of cases—and this was one of them—Khan’s decision was not binding. Either party could still lodge an appeal. The case would then proceed to a legally trained online arbitrator who would handle the case for a further thirty days before issuing a final decision. The costs of the arbitration would be split three ways—between the complainant, the defendant, and the platform company.

Khan was not about to let this case go all the way up to arbitration. Since the amount of money at dispute was not huge, she used a special prerogative that all the virtual judges had: she issued a full compensation to both parties. The client had his money returned, but the supplier was also paid. The platform company took a hit, but it was cheaper than letting the process drag on. Justice—of a sort—was served.

Digital platform companies such as Airbnb, Amazon, Apple, eBay, Google, Uber, and Upwork today engage thousands of people like Khan to handle disputes. EBay alone claimed that it resolved more than 60 million disputes in a single year.² In the same year, the UK court system handled about 4 million cases,³ Chinese courts about 11 million cases,⁴ and US courts about 90 million cases, the majority of which were traffic violations.⁵ In other words, platform companies together probably resolve more disputes now than the entire world’s public courts.

This is not just because people get into more fights on the Internet, though that may be true, too. It simply reflects the fact that so many of our daily interactions are now played out through these platforms. People use platforms to find food, clothes, transport, accommodation, jobs, drugs, entertainment, friends, and even life partners. Firms use platforms to find customers but also real estate, suppliers, contractors, workers, and new innovations. Supervisors use platforms to manage workers, while professors use them to teach. According to one study, 70 percent of industries in the United States are now influenced by digital platforms, either because customers learn about firms and products through platforms or because the entire business from start to finish is carried out through platforms.⁶ If we are lucky and humanity starts taking the climate crisis more seriously, we’ll probably end up spending even more time interacting with the world remotely through digital platforms.

Increasingly, then, the rules that we follow in our daily lives are set by platform companies. They have power over what is permitted and what is prohibited, who can interact with whom, what sorts of agreements are possible, and what kinds of rights and guarantees you have in practice if things go wrong. It’s almost as though they have become some sort of a digital government. Is Microsoft a digital nation and does it have a secretary of state?, asks The Economist.⁷ Apple is basically a small country now, claims The Atlantic.⁸ Who needs a government when you’ve got Amazon to keep things running, quips a columnist at The Guardian.⁹ One Freelance Nation under Upwork, proclaims a blogger.¹⁰

These comparisons reflect the central role that tech firms now play in underpinning and regulating our economic and social activities. An estimated $490 billion’s worth of goods passed through Amazon’s marketplace in 2020¹¹—more than most countries’ entire gross domestic product.¹² The company earned almost $75 billion in fees from merchants who used its marketplace and logistics infrastructure—far more than what most governments earned in tax revenues. The chief executives of leading tech companies are now by many measures more powerful than most countries’ heads of state.

However, not everything is well in these virtual states. Their leaders enjoy immense power without a commensurate level of accountability. Many of the leaders have been discovered abusing that power. Amazon’s managers used their God’s eye view of the market to identify best-selling products, produce copycats, and ensure that consumers bought the copycats instead of the originals. Amazon has taken over the [listing] from me.… Saw me making a profit, and decided to take it from me and sell it themselves, deplored one merchant.¹³ Major platforms have been caught bending the rules of their marketplaces to favor insiders, demanding extortionate fees from those least able to afford them, and stealing lucrative businesses from small entrepreneurs.

How did we end up here? The Internet was supposed to free us from powerful institutions. It was supposed to cut out the middlemen, democratize markets, empower individuals, and birth a new social fabric based on self-organizing networks and communities instead of top-down authority. We will create a civilization of the mind in Cyberspace.… more humane and fair than the world your governments have made before.¹⁴ This is what Silicon Valley’s visionaries promised us. Then they delivered something different—something that looks a lot like government again, except that this time we don’t get to vote. Why did things turn out this way? And what is to be done about it?

THREE THOUSAND YEARS OF HISTORY IN THIRTY YEARS

In each of the chapters in this book I will tell the story of an influential person and iconic platform that helped to shape today’s platform economy, from household names like Jeff Bezos of Amazon to unsung heroes like Kristy Milland of Turker Nation. The stories are based on years of research by my research group at the Oxford Internet Institute and on the work of many other researchers and journalists.¹⁵ Together the stories trace the evolving institutional structure of electronic commerce from 1980s decentralized cyberbazaars to present-day US megaplatforms. The protagonists’ struggles and triumphs illustrate the social and economic forces that shaped today’s platform economy and provide lessons to any who would alter it, whether through political action or program code.

Institutions in the sense used in this book are not buildings or organizations but laws, regulations, traditions, social norms, and other rules of the game that structure people’s interactions in society.¹⁶ A good example is the rule that you must honor any contracts that you sign. The rule is enforced by the state through its courts—and today by tech companies through their dispute resolution centers. It allows people to enter into deals with strangers with some certainty that they won’t get cheated.

In the first part of this book we focus on economic institutions, the rules of the game in the marketplace. Western economic thought used to emphasize the self-organizing nature of markets—how exchanges of goods, services, and labor supposedly emerged from nothing but people’s coinciding needs. But economic sociologists and historians showed that markets also needed institutions like contract enforcement to be viable. Traders in every age needed some means to find each other, understand each other, and trust each other before exchange could take place. What kinds of institutions can we see underpinning markets on the Internet? The nature of economic institutions matters, because they not only determine how efficient markets are, but also how risks and rewards are distributed—who reaps the benefits and who bears the burdens.¹⁷

We start in chapter 2 by examining how exchange worked on the old Internet of screeching modems, before giant platform companies existed. Libertarian cybercowboy John Barlow explicitly rejected the state and other formal institutions as the underpinnings of his new online society. He believed that digital markets could be founded on nothing more than simple reciprocity—the informal rule that you should treat others as you would have them treat you. The rule may have worked in small communities like the one around his ranch in rural Wyoming. But it fell apart as soon as the Internet boom started and electronic communities grew into boom towns.

In chapter 3, we examine another informal institution, reputation—that is, the ancient idea that regard for one’s good name should keep people honest. At the end of the last millennium, Pierre Omidyar built eBay on the basis of this informal institution. He used technology to stretch its capabilities to new levels. But even with modern technology, there were fundamental limits to how much order reputation alone could sustain. To save his project, Omidyar changed tack and turned eBay into a central authority that formally regulated its marketplace—with surprising results.

In chapter 4, we focus on the role of identities in underpinning exchange. Ross Ulbricht wanted to create an online drug market that would offer absolute privacy to its users: not even he as the market’s administrator would know who the users were. But he found that he could not maintain order without some sort of stable identities that linked people’s past actions to future consequences. Like modern states, digital platforms began to assign persistent identifiers to people that the people could not easily leave behind.

In chapter 5, we examine the notion of borders as they pertain to institutions and markets. Straddled on opposite sides of the Atlantic, Odysseas Tsatalos and Stratis Karamanlakis wanted to use the Internet to create a borderless labor market. They constructed virtual institutions that made it possible for people across national boundaries to work with each other as if they were in the same jurisdiction—a platform that is now called Upwork. But to save this new online economy from an economic crisis, they ultimately had to raise their own virtual borders around it.

In chapter 6—the final chapter of the first part—we observe how platform designers began to shift from constructing free markets toward conducting central planning. Uber cofounder Travis Kalanick was a fierce advocate of free-market solutions. Yet insofar as Uber fixed all fares and regulated the numbers of cars on the streets, it was anything but. Thanks to advances in surveillance and information processing technologies, Silicon Valley technologists successfully overcame many of the technical constraints that stymied Soviet planners. Yet fundamental questions about the humaneness of top-down planning remained.

The first part of the book demonstrates that much like historical commerce, electronic commerce started as occasional exchanges with close acquaintances, grew into regular trade structured by personal reputation, and finally multiplied manyfold on impersonal marketplaces secured by organized authorities who controlled entry, kept records of the participants, employed judges to resolve disputes, and enforced rules coercively if necessary. In this sense, the Internet essentially recapitulated the past three thousand years of economic history in thirty years. Now it appears stuck somewhere in the mid-twentieth century, teetering between markets and central planning. The difference is that this time the authorities in charge are not nation states but digital platform companies.

Why did this happen? Why did Silicon Valley technologists end up recreating in digital form the very institutions that they were trying to obsolete? I will address this question in detail in chapter 12. For now it suffices to say that technology ultimately doesn’t change the fundamental social and economic forces that shape how societies are organized. Platform companies don’t appear stately merely by virtue of being powerful but are powerful precisely because in certain important ways they emulate the state.

THE ANCIENT PROBLEM

As our economic interactions moved online, tech companies became the new central authorities that set the rules and protected us against fraud and cybercrime. They provided the institutional underpinnings that Internet commerce needed. But this gave rise to a new problem: how could we hold these digital authorities to account? Though they enabled much trade and prosperity, their leaders were not saints—they were not above using their position to bend the rules and exploit their subjects to advantage themselves and their allies.

The same problem has troubled political philosophers for millennia: quis custodiet ipsos custodes? Authorities protect us, but who will protect us from the authorities? In part II of this book, titled political institutions, we will examine how people in the platform economy have attempted to address this problem. If economic institutions are the rules of the game in the marketplace, then political institutions are the mechanisms through which those rules can be changed and the rulers held to account.¹⁸ We will observe not just ideas and experiments but actual uprisings against the princes of the platform economy.

In chapter 7, we first recall that Amazon founder Jeff Bezos was once hailed as a hero who created an ideal business environment for countless independent merchants. But as soon as Amazon became the dominant marketplace, Bezos turned on his merchants and began extracting extortionate fees and outright stealing lucrative business lines from them. Despite being compared to an autocratic government, Amazon was legally a private corporation. The usual way in which private corporations are disciplined is through market competition: if people don’t like a company’s services, then they are free to vote with their feet and switch to a competitor. But network effects and switching costs meant that it was very difficult for Amazon’s merchants to leave the platform. Competitive pressures turned out to be entirely insufficient to curb platform autocrats’ abuses.

If not through competition, then how else could platform authorities be held to account? In chapter 8, we examine the idea that the whole question could be sidestepped by replacing untrustworthy human authorities with incorruptible machines—solving the ancient problem once and for all with technology. Bitcoin inventor Satoshi Nakamoto and Ethereum cocreator Vitalik Buterin seemed at the cusp of realizing this crypto-anarchists’ dream using so-called blockchain technology. But it became apparent that humans were still needed to write the machines’ rules and to update them whenever circumstances required it. Blockchain may have automated administration but it did not automate legislation. Politics were not eliminated but merely displaced to less scrutable arenas.

In chapter 9, we begin to examine users’ attempts to insert themselves into platforms’ rule-making in the old-fashioned way—through collective action. Kristy Milland and her fellow workers on Amazon’s digital piecework platform Mechanical Turk organized a campaign to try to improve their lamentable working conditions. The workers demanded that Jeff Bezos give them a say in how the platform’s rules were being made. Risks inherent to standing up to the platform prince meant that the campaign failed to amount to any sort of workers’ revolution. But it was nevertheless significant in that platform users were now for the first time claiming a moral right to participate in shaping the laws that governed them.

In chapter 10, we discover another attempt to influence platform rules via collective action. App entrepreneur Andrew Gazdecki mobilized his peers against Apple over a policy change that threatened to destroy his and many other small businesses that depended on the App Store. The difference to Milland and her fellow digital laborers was that Gazdecki and his peers were affluent members of a rising digital middle class who possessed considerable resources and connections. They applied their resources to mount a powerful campaign, and in a rare win for users, the fruit giant yielded.¹⁹

The second part of this book demonstrates that even as our economies have moved online, competition has not been able to ensure that platform companies would treat people and businesses fairly. Why do platforms differ from ordinary companies in this regard? I will address this question in detail in chapter 12. For now it suffices to say that like states, platforms are institutional frameworks, and the choice between alternative institutional frameworks is not an individual choice that can be resolved on a market, but a collective choice. Political institutions for collective decision making are missing from the platform economy, so users find themselves stuck at the princes’ mercies.

Attempts to circumvent the resulting political problem with blockchain technology have missed the mark, because they have focused on decentralizing rule-enforcement, even as the real power lies in rule-making. Meanwhile, people and businesses who depend on platforms for their living have begun to rediscover the tools that toppled past autocrats: resources, alliances, and organizing.

INSPIRATION FROM OUR ANCESTORS

In the final part of this book, we briefly examine the platform economy’s social institutions. By social institutions, I am here referring to institutions whose purpose is to protect and nurture. Modern nation states and especially European welfare states are expected not only to support markets but also to support people through education, hardship, illness, and old age. If platform companies are assuming state-like oversight over markets, are they assuming such oversight over the people who work those markets as well?

In chapter 11, I follow the story of Sofia, an unemployed translator in California. She tries to reskill herself with online courses while paying her bills with gig work but falls seriously ill. A campaign on charity crowdfunding platform GoFundMe is able to stave off financial ruin for her but only for the moment. Sofia’s story illustrates how the platform economy is undermining established social institutions, while the alternatives that platform companies have created still fall far short of the need—to the extent that it is starting to cause problems for the long-term sustainability of the platform empires themselves.

In the final chapter of this book, I draw on all the preceding chapters to explain why digital platforms have become the new virtual states, how they nevertheless differ from our earthly nations, and how we can regain control over them. The whole idea that digital platforms might somehow be commensurate with states seems fanciful. But it would help to explain a lot about the present situation, in which smaller nations feel compelled to appoint tech ambassadors to carry out digital diplomacy with Silicon Valley firms.²⁰ Moreover, it could open up new directions to us as we grapple with the question of how to deal with these companies’ power.

Many social scientists today approach platform giants as a new type of monopolistic capitalist enterprise.²¹ They argue that governments should use competition law to break up platforms into pieces, or public utility law to regulate or even nationalize them. Both approaches were developed at the turn of the twentieth century to fight abusive industrial capitalists. There are signs today that they might not be up to the task of dealing with digital giants.²² Scholars are debating how to update them.²³ But what if these approaches were not up to the task not because they are outdated—but because they are too modern? What if instead of drawing precedent from how our grandparents dealt with capitalists, we ought to be drawing inspiration from how our ancestors dealt with aristocrats? These questions I will explore in the concluding chapter.

Drawing parallels between technology companies and states is admittedly dangerous because it risks legitimizing the companies’ power: the more we hear that tech companies are like states, the more we may come to accept their rule as natural. But if that is true, then such parallels should also help legitimize another idea that we naturally associate with states: that they should ultimately be governed by their people.

HOW TO READ THIS BOOK

The relationship between digital technologies and society is surely one of the big issues of our time after the climate crisis. There is a sense that digital technologies have in some ways begun to undermine our present social order. Political scientists worry that digital election manipulation compromises governments’ legitimacy.²⁴ Security experts warn that cyberweapons challenge states’ ability to protect their citizens.²⁵ Lawyers argue that gig economy apps and cryptocurrencies undermine the legal order.²⁶ In this book I take a slightly different approach to the issue. Instead of asking how technologies are undermining our present social order, I ask how they may be constructing an alternative one.

The chapters in this book are self-contained, and each can be read as an accessible introduction to an influential platform, person, or theory that shaped today’s digital economy. The chapters are presented in a roughly chronological order, starting from the precommercial Internet and ending in the pandemic, so the book can also be read as a basic economic history of Western online commerce. Much of the action takes place on the West Coast of the United States, where the Internet was first created and where most Internet giants still live. But I follow the platforms’ implications to different parts of the world, giving voice to businesses and workers from Europe, Asia, and Sub-Saharan Africa.

If you are thinking about building a digital marketplace yourself and wish to learn about the theory, practice, and pitfalls of doing so, I suggest you start reading from part I. If you are an activist or organizer and wish to learn about the theory and practice of peoples’ struggles against platform companies, make sure to read part II.

If you are drawn to ideas of community and decentralization as alternatives to government authority, I suggest you read chapters 2, 3, and 8. If you are interested in remote work, global development, and labor issues, read chapters 5, 6, and 9. If you are interested in competition policy issues, look into chapters 7, 10, and 12. If you are a busy policy maker, I will forgive you for skipping straight to the final chapter.

By comparing digital platforms with states and other institutions, I am attempting to offer a diagnosis of their power that is broad and fresh and yet avoids the kind of exceptionalism in which everything digital is assumed to be novel. As we shall see, seemingly revolutionary technologies often have surprisingly ancient precedents.

NOTES

1. The details of the dispute resolution process are based on interviews and observations reported in G. Corporaal, S. Windwehr, and V. Lehdonvirta, Impartial or Captured? How Online and Offline Labor Market Intermediaries Handle Workplace Disputes, working paper, University of Oxford. The interviews and observations were collected during organizational fieldwork at the platform company’s premises in Silicon Valley by Gretta Corporaal in the iLabour research project that I led. The details of the case and the agent have been changed to protect participants’ privacy and the confidentiality of the process.

2. E. Katsh and O. Rabinovich-Einy, Digital Justice: Technology and the Internet of Disputes (Oxford University Press, 2017).

3. Judicial and Court Statistics 2010 (UK Ministry of Justice, 2011).

4. 年全国法院司法统计公报 (中华人民共和国最高人民法院, 2010). Annual National Judicial Statistical Communique (Supreme People’s Court of the People’s Republic of China, 2010).

5. Court Statistics Project, State Court Caseload Digest: 2018 Data (National Center for State Courts, 2020).

6. M. Kenney, D. Bearson, and J. Zysman, The Platform Economy Matures: Measuring Pervasiveness and Exploring Power, Socio-Economic Review 19, no. 4 (2021): 1451–1483. See also M. Kenney and J. Zysman, The Rise of the Platform Economy, Issues in Science and Technology 32, no. 3 (2016): 61–69.

7. Schumpeter: The Redmond Doctrine. Lessons from Microsoft’s Corporate Foreign Policy, The Economist, September 12, 2019, https://www.economist.com/business/2019/09/12/the-redmond-doctrine.

8. A. LaFrance, Apple Is Basically a Small Country Now, The Atlantic, February 11, 2015, https://www.theatlantic.com/technology/archive/2015/02/apple-is-basically-a-small-country-now/385385.

9. J. Naughton, Who Needs a Government When You’ve Got Amazon to Keep Things Running?, The Guardian, March 28, 2020, https://www.theguardian.com/commentisfree/2020/mar/28/who-needs-crisis-government-when-youve-got-amazon-coronavirus.

10. N. Todorovic, One Freelance Nation under Upwork, Medium, March 29, 2019, https://nebojsa-todorovic.medium.com/one-freelance-nation-under-upwork-7823984af45c.

11. J. Kaziukėnas, Amazon GMV in 2020, Marketplace Pulse, 2021, https://www.marketplacepulse.com/articles/amazon-gmv-in-2020.

12. World Bank, World Bank National Accounts Data, World Bank Group, 2021, http://data.worldbank.org/indicator/NY.GDP.MKTP.CD?year_high_desc=true.

13. See chapter 7.

14. J. P. Barlow, A Declaration of the Independence of Cyberspace, Electronic Frontier Foundation, 1996, https://www.eff.org/cyberspace-independence.

15. All sources are cited in these notes. See the acknowledgments at the end of the book for more details.

16. D. C. North, Institutions, Journal of Economic Perspectives 5, no. 1 (1991): 97–112.

17. S. Ogilvie, ‘Whatever Is, Is Right’? Economic Institutions in Pre-industrial Europe, Economic History Review 60, no. 4 (2007): 649–684.

18. This notion of a hierarchy of institutions was popularized by economist Daron Acemoglu and his colleagues. D. Acemoglu, S. Johnson, and J. A. Robinson, Institutions as a Fundamental Cause of Long-Run Growth, in Handbook of Economic Growth, ed. P. Aghion and S. N. Durlauf (Elsevier, 2005), 385–472. But it is arguably already latent in the work of economic historian Douglass North and his collaborators. D. C. North, Institutions, Institutional Change and Economic Performance (Cambridge University Press, 1990).

19. I use the term users in this book to refer to both consumers and businesses that use platforms to buy or sell goods, services, or labor. Occasionally, I distinguish business users or professional users from consumers. In economics and management studies, businesses who sell something on platforms are often called complementors. See, e.g., M. A. Cusumano, A. Gawer, and D. B. Yoffie, The Business of Platforms: Strategy in the Age of Digital Competition, Innovation, and Power (HarperBusiness, 2019); J. Rietveld and M. A. Schilling, Platform Competition: A Systematic and Interdisciplinary Review of the Literature, Journal of Management 47, no. 6 (2021): 1528–1563.

20. L. Clarke, Tech Ambassadors Are Redefining Diplomacy for the Digital Era, TechMonitor, February 16, 2021, https://techmonitor.ai/leadership/innovation/tech-ambassadors.

21. N. Srnicek, Platform Capitalism (Polity, 2017); T. Wu, The Curse of Bigness: Antitrust in the New Gilded Age (Columbia Global Reports, 2018); S. Zuboff, The Ageof Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (Public Affairs, 2019).

22. D. Coyle, Platform Dominance: The Shortcomings of Antitrust Policy, in Digital Dominance: The Power of Google, Amazon, Facebook, and Apple, ed. M. Moore and D. Tambini (Oxford University Press, 2018), 50–70; J. Espinoza, EU Struggles to Build Antitrust Case against Amazon, Financial Times, March 11, 2021, https://www.ft.com/content/d5bb5ebb-87ef-4968-8ff5-76b3a215eefc.

23. L. M. Khan, Amazon’s Antitrust Paradox, Yale Law Journal 126, no. 3 (2017): 710–805; L. M. Khan, Amazon: An Infrastructure Service and Its Challenge to Current Antitrust Law, in Digital Dominance: The Power of Google, Amazon, Facebook, and Apple, ed. M. Moore and D. Tambini (Oxford University Press, 2018), 98–132.

24. D. Tambini, Social Media Power and Election Legitimacy, in Digital Dominance: The Power of Google, Amazon, Facebook, and Apple, ed. D. Tambini and M. Moore (Oxford University Press, 2018), 265–293.

25. L. Kello, The Virtual Weapon and International Order (Yale University Press, 2019).

26. J. Prassl, Humans as a Service: The Promise and Perils of Work in the Gig Economy (Oxford University Press, 2018); K. A. Thelen, Regulating Uber: The Politics of the Platform Economy in Europe and the United States, Perspectives on Politics 16, no. 4 (2018): 938–953.

I

ECONOMIC INSTITUTIONS

2

RECIPROCITY: THE GOLDEN RULE IN CYBERSPACE

Therefore, live with self-restraint and pay your best attention to dharma, and treat others as you treat yourself.

—Mahābhārata Shānti-Parva 167:9

When