tech companies have also embraced the power narrative that regulation will crowd out innovation
Introduction: The Battle
Those who praised the democratizing possibilities of technology and social media platforms failed to appreciate that repressive authoritarian regimes could be tech-savvy too.
Instead this book begins from the premise that these incidents point to systemic problems that need unpacking: the fact that our social, professional, and civil lives are increasingly digitized and, essentially, all aspects of digitization are in the hands of private companies; that certain technologies have inherently antidemocratic characteristics, while laws to protect democratic values and the rule of law are lagging; and that, most important, democratic governments’ outsourcing of key functions has led to a hollowing out of governments’ core capabilities. These systematic problems are now undermining the core principles of democracy: free and fair elections, the rule of law, the separation of powers, a well-informed public debate, national security and the protection of civil liberties such as freedom of expression, the presumption of innocence, and the right to privacy.
Sam Altman’s Worldcoin, for example, aspires to build a global identity database by asking people in developing countries to scan their irises, in return for a bit of cryptocurrency;
Egypt, for example, has relied heavily on Chinese investment to modernize its telecommunications infrastructure and even construct a new smart city. The North African country has now also adopted China’s model of internet governance via a new cybercrime law, and Egyptian government officials routinely attend Beijing’s censorship training.
1. The Code
After all, in my experience, tech industry leaders mostly cared about policy when they tried to prevent it from having an impact on their businesses.
Conversations then continued informally under what are known as the Chatham House rules, according to which attendees can later relate the content of the discussion but are forbidden from attributing specific comments to individual speakers. This rule, which is meant to encourage frank conversation, gave these titans of industry a chance to break from their polished talking points and to express themselves frankly, including their true distaste for oversight of any kind.
I quickly found myself cornered. Questions and critiques of Europe’s relationship with technology were firing at me from all sides. Did Europeans realize their tendency to overregulate was the reason why no equivalent of Silicon Valley existed there? Was jealousy to blame for a vindictive zeal to obstruct American tech firms? Why were Europeans so emotional about privacy protections? Didn’t the behavior of teenagers posting their entire lives online, including those in Europe, indicate that privacy was already dead?
These tech leaders easily dismissed concerns about pressure on fundamental rights as “emotional.” They were proud of the corporate and political power they wielded to influence society’s technological future— and for building an industry that, until then, had largely operated outside the grasp of regulators.
But it did not take long before companies were trying to wall off and brand the online experience. Back in 1996, Berners-Lee cautioned, “The web has exploded because it is open. It has developed so rapidly because the creative forces of thousands of companies are building on the same platform. Binding oneself to one company means one is limiting one’s future to the innovations that one company can provide.”
Today’s industry is a far cry from Berners-Lee and Barlow’s vision of emancipation through the wide availability of technology. Market pathologies have replaced the traditional institutions of the past with a new generation of powerful gatekeepers and elites. Despite their trendier branding, tech oligarchs are just as power hungry—and possibly more powerful—than the government institutions that internet pioneers sought to challenge decades ago.
A later story in the New York Times featuring four courageous whistleblowers indicated that Greyball was being used on a global scale. 21 The sheer audacity of Greyball showed that Uber and its peer companies believed they could act with impunity. The tech firms not only objected to specific regulations but thought they were above the entire regulatory regime.
Tech companies have been so emboldened to flout regulation and defy governments because governments need them. In direct contrast to the anarchist and libertarian dreams of the internet’s founders, Big Tech has become an agent of the state. Tech Inquiry has reported that agencies like the Federal Bureau of Investigation (FBI), the Federal Bureau of Prisons, Immigration and Customs Enforcement (ICE), the U.S. Department of Defense, and the U.S. Drug Enforcement Administration have thousands of deals with Amazon, Dell, Facebook, Google, HP, and IBM. 23 Since 2016, Microsoft alone has had over 5,000 subcontracts with the Department of Defense. Amazon and Google are next in terms of volume, with 350 and 250 contracts, respectively. 24
Beyond covert resistance to government regulation, as exemplified in Uber’s Greyball, tech companies have also embraced the power narrative that regulation will crowd out innovation. In my time as a member of the European Parliament, I heard this Silicon Valley mantra echoed by almost every tech lobbyist. Without exception, any proposed piece of legislation would be presented as an undue, old-fashioned, and harmful intervention that risked stifling innovation that would surely benefit society. If Europe was to have a chance at producing a success remotely like that of Silicon Valley, lawmakers had better back off.
Today it is used for mapping and navigating by companies and individuals across the world, yet the tech sector wants us to believe that it has simply pulled itself up by its own bootstraps.
Section 230 of the CDA has fundamentally shaped the internet as we know it today, particularly by offering limitations to the liability of online platforms. The original idea—that selling a stolen bike or handbag on eBay should lead to liability for just the thief, not the platform—has morphed into an immense abdication of responsibility for content including hate speech, conspiracies, and discriminatory language.
However, Bush’s most important tech legacy was the introduction of extraordinary surveillance programs in the name of fighting terror after 9/11. Along with Vice President Dick Cheney and Attorney General John Ashcroft, he defended warrantless wiretapping and created the Department of Homeland Security, which would also step up cybersecurity capacity. With its attention focused on issues of terrorism, it is no surprise that Ed Black, a corporate tech lobbyist, has described the Bush administration as largely being “AWOL on technology policy.” 41 By neglecting the tech industry, the Bush administration allowed the new generation of tech firms to build the foundation for the sector’s future dominance. At the same time, the tech industry realized that it could achieve favorable policy outcomes by making itself useful to the state. Silicon Valley, which first emerged in the post-World War II military boom, carefully integrated itself into the rapidly expanding U.S. national security regime during the War on Terror.
President Obama’s “war room” often included high-level folks from the Silicon Valley elite. Google’s CEO, Eric Schmidt, supported the president as a campaign adviser and fundraiser.
In 2013, for example, the FTC settled with Google after a two-year investigation into its control of standardized technology patents and dominance in the search engine and adtech markets. 45 Many continue to see the settlement as a missed opportunity to ensure more competition in the digital economy. At the time, the FTC sympathized with Google’s argument that low competition was a necessary evil to ensure that consumers could get “better, faster, more valuable answers to their queries.” 46 This logic and precedent, however, laid the groundwork for sweeping market consolidations by companies like Amazon and Apple who similarly justified their outsize influence with appeals to consumer experience.
As in most areas, President Donald Trump handled the issue differently from his predecessors. He used social media as an unprecedented loudspeaker. Disinformation and conspiracy theories were shared on White House letterhead and from the West Wing podium. Only when President Trump and his allies saw their accounts or posts removed for inciting violence did they begin to talk about reining in social media giants. And even then, there was little thought or effort beyond political expediency.
Though the decision to leave crucial decisions to private companies could be explained by either ignorance of the industry or by blind faith in the free market or both, there is a more proximate explanation: this policy was supported by a wide-ranging cast of public officials who left office to work for tech companies— and, sometimes, even came back to government afterward.
Every major tech company has begun hiring former government officials, bringing knowledge about politics and policy in-house, while building valuable connections to key policymakers.
Even Eric Holder, the former attorney general under President Obama who had spent his entire career in public service, went to work for Airbnb to help the company with its antidiscrimination policies. 66
2. The Stack
China itself plays a different but critical role in the semiconductor supply chain as the supplier of rare earth materials. At the moment, the country has a near monopoly on several of the materials critical for microchips, including lanthanides, yttrium, and scandium, to name just a few. Tensions between China and the EU, the United States, or Taiwan thus pose an existential risk to the tech supply chain. Beijing could close access to its mines, leaving the world without the basic components needed to make most high-tech goods.
The company, located in the Netherlands, would only be able to do so after obtaining a license from the government. After the U.S. government expressed its concerns and shared classified intelligence reports with the Dutch government, the license was revoked. Though few details about this ordeal have been disclosed by ASML, the Dutch government, or the U.S. government, the reports presumably warned that the deal would allow China to make critical advances. Similar pressure from Washington has prevented exports by ASML to China, as recently as January 2024. 15
Today, however, they are a billion-dollar industry of their own. 28 Since 2012, U.S. companies, including Alphabet, Amazon, Meta, and Microsoft, have significantly invested in undersea cables linking both European and non-European Mediterranean states to other parts of the world, such as countries in Africa and Asia. By 2024, the companies are expected to own a stake in more than thirty long-distance undersea cables connecting every continent. 29 As a result, their combined share of the world’s total cable capacity has increased from 10 percent to 66 percent in a decade. China Mobile and Meta have laid undersea cables in a partnership, despite tensions between the U.S. and Chinese governments. 30 The privatization of this global critical technology infrastructure further concentrates power in the hands of technology companies, but it does so in a way that is mostly invisible to citizens and civil society. The companies seem to prefer it this way, as they generally offer little information about their ownership of undersea cables.
As the common saying goes, innovation is often born out of necessity. Those who live in the Netherlands know that all too well, as it is a country that would largely be underwater had it not been for centuries of innovation in methods of defending the land against sea and river flooding. The city center of Amsterdam lies two meters (about 6.5 feet) under sea level today, and much of the country would flood without an ingenious system of dikes and other water defenses. To reclaim land and expand the country, the Dutch have also become experts in dredging, an age-old practice that is used for anything from building or deepening harbors, digging canals, or developing artificial islands. As my compatriots are fond of saying, “God created the world, but the Dutch created the Netherlands.”
Polders, the land sitting between two dikes, have become characteristic features of Dutch landscapes and represent another innovative way of managing flows of water— a way that forced people with divergent interests to interact and collaborate.
The data center debate in Zeewolde raises the following question: Are poorly staffed and resourced local governments equipped to endure the sophisticated tactics and lobbyists employed by large tech firms when seeking licenses and approvals? The local Zeewolde representatives perform their function part-time, and the decision concerning the Meta data center alone generated tens of thousands of pages to work through.
But after buildings are put in place, data centers end up employing relatively few people. Even though the U.S. government provided tax breaks as high as $150 million for Apple in Iowa, only fifty permanent jobs in the center were created. 36
As of now, the data center skeptics seem to have succeeded. EirGrid announced in spring 2022 that it was pulling the plug on thirty Irish data center projects due to electricity curbs, a major reversal for the provider and the Irish government. 53
3. The Weaponization of Everything
DarkSide claims the attack was just about money: “Our goal is to make money and not creating problems for society. We do not participate in geopolitics, no need to tie us with a defined government and look for our motives.”
The group apparently has some limits and avoids attacking hospitals and schools. The reason may be quite self-serving, however: these public organizations often won’t have the resources to pay up. The group also seems to avoid areas where allies of the Russian government reside.
Yes, a T-shirt: the NCSC played into the sentiment of pride and competition between hackers and gave out T-shirts at a hackathon that said, “I hacked the Dutch government, and all I got was this lousy t-shirt.”
For the moment, it is safe to say that cybersecurity is more of a business model than a policy model. That increases the risk of moral hazard, with companies reaping profits from the sale of their products but cases of failure falling onto the public to absorb.
So far the digital encroachment on national security has not been accompanied by a transformation in law to oversee the operations of these now-critical security operations in the private sector. The lagging application of legal principles, combined with the typical opacity of corporate information, means that transparency and accountability in the public interest are rare. Public authorities ranging from water to energy companies, local police forces to hospitals, are no longer able to run and protect their systems without hiring cybersecurity companies— and they frequently do not know what the cybersecurity firms are even doing.
4. The End of the Public Interest
Most shockingly, reporting from the New York Post found that rogue New York Police Department officers had Clearview installed on their personal devices, even after the department’s own facial recognition unit passed on Clearview due to a variety of concerns.
A group of over two hundred Palantir employees— mostly data scientists and software engineers— were demanding that the company stop providing tools to ICE, the federal agency tasked with managing immigration to the United States and protecting America’s borders. While Palantir had worked for ICE for years, working for ICE under the administration of President Donald Trump, which had been publicly criticized for its family separation and detention tactics, felt morally beyond the pale to the letter’s signatories. ICE had started showing up in neighborhoods and offices unannounced and rounding up undocumented immigrants who had been living in the United States for decades. These new tactics were the agency’s attempt to fulfill Trump’s promise to deport “millions of illegal aliens” using a “deportation force.” 75 Palantir’s data-slicing products were critical for helping ICE find their targets, and employees were horrified at how the products they had built were being used.
It was built on the collective skills of computer scientists, data crunchers, and intelligence officers with initial funding from In-Q-Tel, the venture capital branch of the U.S. Central Intelligence Agency. 79 Palantir has since penetrated governmental operations in a way few other companies have. It works for the U.S. Food and Drug Administration (FDA) to “modernize the food supply chain,” helps the U.S. Army analyze data to gain battlefield advantages, and develops AI for the U.S. Special Forces.
5. Tech on the Front Lines
Days before elections took place in 2020, Apple and Google dealt a significant blow to Russian voters by removing the app that Navalny’s associates developed to help identify the strongest candidate opposing Vladimir Putin in each of the 225 voter districts.
Like many autocratic regimes, Russian authorities pointed to national security as a justification for their censorship. According to the official party line, Navalny’s movement was an “extremist” and “terrorist” organization. 3 U.S. tech companies could have rejected the absurdity of Russia’s claims, but the authorities raised the stakes. They threatened to hold Apple’s and Google’s employees in Russia personally accountable for the decisions made in the company’s American headquarters. The threat was effective: the companies cited the danger to employees as their main motivation for removing the app from their platforms.
Ronan Farrow has called Elon Musk “a private citizen with a private company who had become the arbiter of the outcome of this war.” 46 There are few people in the world not in government of whom the same can be said.
Most famously, after World War II, states signed the UN Charter, which held them to a host of commitments regarding human rights and the use of military force. Additional treaties and conventions bind signatory governments to uphold everything from the laws of armed conflict to trade agreements to the laws of the seas. Typically these treaties also include mechanisms of accountability in case signatories don’t comply.
As far as we know, tech companies have abided by international conventions in the case of Ukraine, but there is no reason to expect that they will be on the right side of history or human rights law in the next conflict. Elon Musk has recognized that in the current geopolitical climate, standing by Ukraine is good for business; Russia is a relatively limited tech market, and catering too much to the Kremlin might carry commercial consequences for these companies in the United States. But the market calculus changes with respect to China and some Middle Eastern nations. In conflicts involving those countries, companies may prioritize their business interests over the greater good.
Notably, companies are playing an ever more critical role in this strange cyber dividing line between war and peace. Microsoft takes down state-sponsored malware across the world and refers to court orders to legitimize its actions. Google’s Threat Analysis Group, one of the company’s elite security operations, terminates YouTube channels for running coordinated influence operations and shuts down countless phishing operations targeting high-level officials and journalists. 56 Amazon and its web hosting platform AWS have sophisticated machine learning algorithms geared toward catching state-sponsored threats and accompanying suspicious log-ins. 57 Much of this sounds like cyberdefense and cyberwarfare, but it’s often taking place between a country and a private company, not between two sovereign nations.
Yet there are no elections for consumers to share thoughts on corporate policy; CEOs cannot be voted in (or out) by the public; C-SPAN doesn’t cover these companies’ internal deliberative processes. The decisions that they make in the public interest are locked behind the fortress of private-sector protections. And unless democracies begin to claw back their power from such companies, they will continue to experience the erosion of their sovereign power.
6. The Framers
Cambridge Analytica’s winning streak screeched to a halt when data scientist Christopher Wylie, who had worked at the firm until 2014, decided to blow the whistle on the firm’s practices.
“We’ll give you total freedom. Experiment. Come and test out all your crazy ideas,” he says the company told him. 2 The offer was too tempting to refuse, so he joined and helped build systems that could be used to manipulate and nudge citizens into voting for conservative candidates, often pushing them down rabbit holes of wild conspiracy theories online. “One of the things that I learned about myself is that I can get involved in an idea and forget, in a really terrible way, what are the actual consequences of what I’m working on,” Wylie later recalled. 3
By 2017 Wylie could no longer stand what he had built and decided to go public. He contacted the journalist Carole Cadwalladr at the Observer and explained to her how Cambridge Analytica surreptitiously acquired the personal Facebook data of fifty million people against the private platform’s rules. The firm then combined that data with other datasets to create the targeting engine for the influence machine he built. It was a “full propaganda machine.” 5 Or, as Wylie put it more glibly, it was “Steve Bannon’s psychological warfare tool,” referring to one of Trump’s senior advisers.
This tactic, called framing, has been deployed by Silicon Valley companies for years to avoid scrutiny. Framing allows tech leaders to present innovation and technological change in clever and favorable ways while constantly warning about what would happen if overbearing regulation were enacted. Ultimately, they have long understood that perception is the crux of political debate and lobbying. So before we discuss tech’s influence and lobbying operations, we should first understand how they use framing to present their antiregulation message.
In the book Don’t Think of an Elephant!, George Lakoff, a longtime professor of cognitive science and linguistics at University of California, Berkeley, explains how the success or failure of a political agenda can depend on how it is framed.
The language we use to describe ourselves and policies, Lakoff argues, is often the deciding factor for how we view issues and problems. In some cases, language prompts our brains to think of things even when we’re explicitly told the opposite.
Antiabortion advocates have cleverly framed themselves as pro-life. Who, after all, is against life? This framing has forced abortion advocates into an unfortunate defensive position. They call themselves pro-choice, but even though that is technically accurate— they support women having the ability to make decisions about their body— it is a less powerful linguistic frame.
“I am not a crook.” No one remembered the point he was trying to get across; they remembered the word “crook,” and that five-word phrase helped end his presidency. 21
The second frame incessantly pushed by tech leaders and lobbyists is that “regulation stifles innovation.” This frame is particularly pernicious because it creates a false dichotomy between regulation and innovation, treating them as mutually exclusive phenomena. In fact, regulation often spurs innovation. For instance, much of the increase in cars’ fuel efficiency over the past few decades has been driven by higher standards enforced by government transportation agencies like the U.S. Department of Transportation. 22
By leveraging framing— and benefiting from the framing of other talking heads— tech companies have effectively shaped people’s understanding of the value of technology products and the effectiveness of regulation. As a result, the entire debate is playing out with their terminology.
What the RFOB lacks in dollars it makes up for in integrity. Yet this group of tech experts, journalists, and civil rights leaders is not and would never claim to be a substitute for enforceable law.
At first glance, these agreements seem like good corporate behavior and good-natured cooperation between the public and private sectors. In reality, they end up risking to replace legal requirements with soft, nonbinding “promises.”
In January 2020, Microsoft became the first company to open an office to be represented at the UN. It was a surprising move from a company that has said that “we’re not a government, we don’t pretend to be.” 55 But Microsoft’s reasoning for opening the office makes sense. When describing the move, a Microsoft leader enthused, “The team is looking to deepen its work with the UN and identify new projects to advance global multi-stakeholder action on key technology, environmental, humanitarian, development and security goals as well as helping to advance the UN Sustainable Development Goals.” 56 In other words, Microsoft wanted to be at the diplomatic negotiating table.
Microsoft has perfected lobbying, the art of influencing without being seen. But it’s a tactic used by nearly all major tech companies. On top of using clever framing and announcing a whole host of toothless agreements, technology companies have excelled at lobbying governments to influence and prevent regulation. These efforts are aided by the companies’ massive bank accounts. In 2021 Amazon, Apple, Facebook, and Google collectively spent $55 million on lobbying the U.S. federal government—more than any other industry. 60 In 2022 they went even further, spending almost $70 million, a record-breaking amount. 61 Apple alone increased its lobbying spending by 44 percent. It’s the same story for the EU: in 2021 each of the major tech companies increased its lobbying spending in Brussels. 62
Together with delivery companies, online taxi services spent more than $200 million backing Proposition 22, making it the most expensively lobbied proposition in California history. The companies also pushed messages to users through pop-up messages in their own apps, warning that they would have to wait longer or pay more if the proposition failed. 63 Those opposed to the proposition were simply outgunned— as they could not match the companies’ spending nor access people’s phones and data in the same way. 64 In the end, the companies prevailed, and 58 percent of voters decided to retain the designation of independent contractor for drivers and other gig workers.
In the summer of 2022, I was up in the wee hours of the morning to feed my baby daughter. I did what every sleep expert cautions against: I opened my phone, started flipping through social media, and eventually opened my inbox. I
The OpenAI saga rivaled an Agatha Christie novel in its intrigue. It is also a case study for the thesis of this book: absolute corporate power corrupts absolutely. A nonprofit board, whether or not its initial decision was sound, was bullied into reinstating Altman by OpenAI’s biggest corporate backer, Microsoft.
According to an old Dutch saying, “The butcher should not test his own meat,” which is the equivalent of saying, “You should not grade your own homework.” To date, tech companies and democratic governments have largely ignored this wisdom, as both relied on forms of self-regulation— oversight councils, review committees, company commitments, industry-wide agreements, declarations of principle, ethics statements, multistakeholder processes, and ad hoc measures— that have distracted from the task of adopting laws and organizing their enforcement.
7. Reclaiming Sovereignty
On the day-to-day side of regulation, there are also targeted efforts to address specific products, services, and technical challenges. In 2017, EU regulators ended roaming charges for Europeans using their devices in other EU member states, cutting phone bills for tourists, business travelers, and students going to college abroad. Similarly, the EU has adopted a common charger for many electronic devices. From 2024, rules will require all handheld devices, including millions of cell phones, tablets, digital cameras, headsets, and video game consoles, to use a single uniform charging outlet. In 2026, laptops will be added to the list of products. Simple commonsense rules like this make life easier for everyone, make technology more affordable, and cut down on unnecessary waste.
Elsewhere in the Capitol, a bipartisan group of U.S. senators introduced the Restricting the Emergence of Security Threats That Risk Information and Communications Technology Act (RESTRICT Act). As a lawmaker in Europe, I was always amazed by how my American peers excelled at conjuring convenient and complicated abbreviations as they came up with new names for laws. The RESTRICT Act, which Ohio senator J.D. Vance called the “PATRIOT Act for the digital age,” would allow the secretary of commerce to restrict or ban digital products like apps from America’s foreign adversaries, including China, Cuba, Iran, North Korea, Russia, and Venezuela. 34
People like Lina Khan, who chairs the FTC, are working from the view that market abuses are detrimental to both economic competition and the democratic process. The FTC is one of the two main government bodies in the United States responsible for implementing antitrust regulation, as well as protecting consumers from corporate harms. During her confirmation hearing before the Senate, Khan defended the idea that antitrust regulation should “protect our economy and our democracy from unchecked monopoly power.” 37 Amazon and Meta worked diligently (but unsuccessfully) to prevent Khan from being appointed to the role, which underscores how determined she is to take on these companies. 38 Since her appointment, Khan has worked to expand the FTC’s mandate and address the full range of anticompetitive and predatory behaviors that tech companies have employed. For example, the FTC has filed an antitrust lawsuit against Amazon over its allegedly anticompetitive practices and plans to update its settlement with Meta to bar the company from monetizing teens’ data. 39
Paradoxically, the EU and the United States are mirror images of each other when it comes to regulating technology. The EU has passed sweeping regulation with an emphasis on civil liberties, but in the absence of a strong industrial base its actions have not led to the development of alternative products with strong EU values baked into them. Meanwhile, the few U.S. rules that do come to fruition are anchored in national security and have global implications. If the two regions combined their approaches, they could offer the best of both to the world: the EU borrowing some geopolitical realism and the United States embracing the need for protecting civil liberties as much as national security.
China is home to over one billion internet users, as well as some of today’s most powerful technology companies. 61 Over the past twenty-five years, autocrats in Beijing have systematically ensured that the country’s technological revolution has not supported democratization at home— instead serving as an instrument of the communist system. The CCP has built a system in which censorship, surveillance, and propaganda are essential: troves of data are harvested to train new applications of AI but are also used to repress the Uyghur Muslim minority. The population, significant numbers of whom are imprisoned in “reeducation camps,” are used for testing new technologies, such as emotion recognition software. 62 Extreme surveillance registers and controls their everyday movement through a government app they are forced to use. AI is used to track even the most minute changes in people’s faces, including in the pores of their skin. Uyghurs are also forced to share their biometric data— namely, DNA samples. 63 Darren Byler of Simon Fraser University has bleakly observed that “Uyghur life is now about generating data.” 64
The most popular app in the country, Study the Strong Country (which is also translated, perhaps more accurately, as “Study Xi to Make the Country Strong”) inculcates Chinese netizens with Xi Jinping’s musings.
In 2019, during the height of demonstrations against the looming extraditions to China from Hong Kong, Apple removed HKMapLive from its app store. The app was meant to help demonstrators track the movements of the police in real time. Earlier Apple had removed the Quartz news app because it contained “content illegal in China.” 73 The company also took down China’s most popular Quran app, which millions of Muslims use. 74 Tens of thousands of other apps have been removed on orders from the CCP, including foreign media apps, gay dating services, services that help encrypt data, and apps with information about the Dalai Lama and Tibet. 75 Rebecca MacKinnon, who previously ran CNN’s Beijing bureau and serves as the vice president for global advocacy at Wikimedia, has called app stores the “new choke point of free expression.” 76 In our connected world, companies have the power to grant or withhold freedom of the press, freedom of expression, and free assembly.
For instance, after a mass shooting in San Bernardino, California, Apple refused an FBI request to break the encryption of the iPhone of one of the two suspects, and yet it is now caving to China’s demands.
Most notably, the Chinese government has been willing to forsake economic gains to maintain its political grip. The most infamous example is the CCP’s intervention to prevent the initial public offering of Ant Group, part of business titan Jack Ma’s Alibaba retail empire, and a subsequent crackdown on the company. The intervention was allegedly caused by an investigation that found that the offering would financially benefit political families in China who could challenge Xi’s leadership. 82 After the intervention, Ma spent time outside of China in the Netherlands, Japan, and Spain and did not speak publicly for almost two years, during which time he instead took up painting and collecting art. 83
As the Financial Times has reported, “Supporters argue that India has found a world- beating solution for building out and regulating the online commons that is more equitable than the US’s laissez-faire approach, more innovative than the EU’s regulation- heavy model and more transparent than China’s totalitarian template.” 107 Some tech leaders, like Satya Nadella, the Indian- born CEO of Microsoft, have praised the policy, impressed by the government’s digital fluency.
But operating in India means abiding by the government’s often strict dictates. One of the most far reaching is an amendment to the IT law that forbids online platforms from allowing the posting, hosting, or sharing of “misleading information” about the government. 121 YouTube complied with 1.2 million government takedown requests in the first quarter of 2022 alone, the highest number of any country in the world.
8. Prioritizing the Public
The exact minutiae of machine learning models constantly change, but the broad domains that AI impacts— the economy, health care, and the online information ecosystem— are far more enduring. In a 2023 Euronews interview, European Commissioner Thierry Breton clarified that this law did not intend to merely regulate AI but rather “to organise our digital space.” 1 The AI law would be systemic, not hyperspecific. To enact their vision, lawmakers adopted an approach frequently used in product liability, which regulates everything from vehicles to pharmaceuticals. The landmark law that was adopted in March 2024 identifies the level of risk associated with various AI applications and offers mitigating measures based on this classification— just as the European Drug Agency and the U.S. FDA distinguish controlled substances from less addictive medications.
What follows in this chapter are new concepts and practical solutions to achieve just that. The initiatives proposed prioritize the public interest, cut irresponsible dependencies, and vow to reverse the tech coup— tech companies’ takeover of governance, politics, and policy— which I have discussed throughout this book.
Just as we’re skeptical toward medical studies conducted by pharmaceutical companies, we shouldn’t embrace the confident assessments of researchers who are on the company payroll.
As Ben Read of cybersecurity company Mandiant reminds us, with misinformation, “you can build uncertainty, confusion, and distrust. It doesn’t need to stand up to a close reading to have some effect on the population; it erodes trust in all messages.”
I was well aware that, drawing on their legal expertise, they would add critical suggestions to make sure the law realized in practice what I intended. Parliaments typically have such teams of independent legal experts available, as most members are not trained in law. Without expert scrutiny, many proposals that parliamentarians dream up would never survive their first legal challenge.
In this spirit, the creation of a technological expert service should bring independent technological expertise into the legislative process. After all, elected officials cannot all be expected to be experts in the detailed working of technologies.
Currently, an elected official’s lack of knowledge is too easily filled in by the cleverly framed talking points of corporate lobbyists. Moreover, technical details, when they are not adequately anchored in law, can be used to circumvent the spirit of the law.
Resources are available: the Dutch government alone spends about 80 billion euros a year on procuring technologies, on behalf of seventeen million people, less than the population of the state of New York.
Conclusion: Stop the Tech Coup, Save Democracy
When tens of millions of jobs end up displaced by AI, shareholders of companies will benefit, while the price of unemployment or reeducation will fall to societies to shoulder.