Information Technology and Democracy: a Widening Gulf

Information Technology and Democracy: a Widening Gulf

"Surveillance capitalism is...a coup from above...an overthrow of the people's sovereignty and a prominent force in the perilous drift towards democratic deconsolidation..." — Shoshana Zuboff, The Age of Surveillance Capitalism^[1], 2019

“We are being lied to...told that technology takes our jobs, reduces our wages, increases inequality, threatens our health, ruins the environment, degrades our society, corrupts our children, impairs our humanity, threatens our future, and is ever on the verge of ruining everything.” — Marc Andreessen, "The Techno-Optimist Manifesto"^[2], 2023

Anxiety over technology and geopolitics is pervasive today. Yet there is a more fundamental conflict underway than that between great powers over technical supremacy. More deeply, the path technology and democracy as systems have taken have put them at loggerheads and the ensuing battle has claimed victims on both sides.

The dominant trends in technology in recent decades have been artificial intelligence and blockchains. These have, respectively, empowered centralized top-down control and turbo-charged atomized polarization and financial capitalism. Both outcomes are corrosive to the values of democratic pluralism. It is little surprise, then, that technology is widely seen as one of the greatest threats to democracy and as a powerful tool for both external authoritarians and those who would subvert democracies from within.

At the same time, democracy was once a radical experiment to scale the governance of a city-state to many millions of citizens spread across continents. A quote on the wall of the memorial in Washington, D.C. to United States Founding Father Thomas Jefferson reads "(L)aws and institutions must go hand in hand with the progress of the human mind... We might as well require a man to wear still the coat which fitted him when a boy as civilized society to remain ever under the regimen of their barbarous ancestors." Yet today democracy has become a synonym in much of the world for the increasingly desperate effort to preserve rigid, outmoded, polarized, paralyzed, and increasingly illegitimate governments. We should not be shocked, therefore, at the disdain that so many technologists have for democratic participation, viewing it as an impediment to progress, nor should we be surprised by the fear among so many advocates of democracy that technical advance will result in the dominance of authoritarian adversaries or internal collapse.

In this book, we hope to show that this tragic conflict is avoidable and that, properly conceived, technology and democracy can be powerful and natural allies. However, it is no accident that arguments in this direction evoke eye-rolling in many quarters. A gulf of grievance and distrust between the two sides of this divide has developed over the last decade and will not easily be laid to rest. Only by fully acknowledging and embracing the legitimate concerns and critiques of both sides of this conflict shall we have a chance to see its root cause and seek to transcend it. Thus, we begin by drawing out these grievances with a generous spirit, accepting critiques that have raised broad concerns even when they are imperfectly supported by the available evidence. Trying to reconcile these extreme divergences offers an opportunity to raise the ambition of democratic technology.

Technology’s attack on democracy

The last decade of information technology has threatened democracy in two related yet opposite ways. As Daron Acemoglu and James A. Robinson famously argued, free democratic societies exist in a “narrow corridor” between social collapse and authoritarianism^[3]. From both sides, information technologies seem to be narrowing the corridor, squeezing the possibility of a free society.

On the one hand, technologies (e.g., social media, cryptography and some other financial technology) are seen to be breaking down the social fabric, heightening polarization, eroding norms, undermining law enforcement and accelerating the speed and expanding the reach of financial markets to the point where they are unaccountable to democratic polities. We shall call these threats “anti-social”. On the other hand, technologies (e.g., machine learning, foundation models, the internet of things) are increasing the capacity for centralized surveillance, the ability of small groups of engineers to set patterns in systems that shape the rules of social life for billions of citizens and customers and reduce the scope for people to meaningfully participate in shaping their lives and communities. We will call these threats “centralizing”. Both threats strike at the heart of democracy, which, as Alexis de Tocqueville famously highlighted in Democracy in America, depends on deep and diverse, non-market, decentralized social and civil connections to thrive^[4].

The antisocial threat from recent technologies has social, economic, legal, political and existential faces.

• Socially, there is growing evidence that while social media have offered powerful new platforms for those who have previously been socially isolated (e.g. sexual or religious minorities in conservative locales) to forge connections, on average these tools have contributed to exacerbating social isolation and feelings of exclusion^[5].

• Economically, the geographic, temporal and multiemployer flexibility facilitated by the internet and increasingly by telecommuting have expanded opportunities for many workers in developing countries or who fit poorly in traditional labor markets. Yet they have largely been unmatched by the emergence of appropriate labor market institutions (such as unions and labor regulations) that allow workers to share in the potential benefits of these arrangements. Thus, they have tended to raise workplace precarity and contribute to the “hollowing-out” of the middle class in many developed countries^[6].

• Politically, polarization and the influence of extremist parties has been steadily rising in many developed democracies. While the role of the internet-based social media landscape is a topic of significant academic debate, recent surveys suggest that these tools have fallen far short of their promise of strengthening social and political bonds across differences and may well have contributed to the secular rise in polarization since 2000, especially in the US^[7].

• Legally, the proliferation of financial innovation in the past few decades has led to limited measurable consumer benefits (in terms of risk reduction, capital allocation or access to credit) while increasing risk in much of the financial system and proliferating financial instruments, thereby challenging or even skirting existing regulatory regimes intended to mitigate these harms^[8]. While innovations surrounding housing finance leading up to the 2008 financial crisis were some of the most important examples, perhaps the most extreme (if more contained) case has been the recent activity around digital “crypto” assets and currencies. Given their mismatch for existing regulatory regimes, they have offered pervasive opportunities for speculation, gambling, fraud, regulatory and tax evasion, and other anti-social activities^[9].

• Existentially, there is growing concern that the fragmentation of the social sense-making and collective action capacity is dangerous in the face of the increasing sophistication of technologies of mass destruction with impact ranging from environmental devastation (e.g., climate change, biodiversity loss, ocean acidification) to the potentially apocalyptic disruptions of more direct weapons (e.g., misaligned artificial intelligence and bioweapons)^[10].

Yet even as technology is seen to erode the cohesion of democratic societies, it is also increasingly seen to threaten democracy by strengthening the control of governments and centralizing power in the hands of a small group of private actors.

• Socially, perhaps the most consistent effect of information technology has been to expand the availability and accelerate the spread of information. This has dramatically eroded the sphere of private life, making an increasing range of information publicly available. While such transparency might in principle have a range of social effects, the power to process and make sense of such information has increasingly concentrated in the hands of corporations and firms that have a combination of privileged access to the information and the capital to invest in large scale statistical models (viz. “AI”) to make these data actionable. Furthermore, because these models improve greatly with access to more data and capital, societies where central actors have access to very large pools of both have tended to pull ahead in the perceived “AI race”, putting pressure on all societies to allow such concentration of informational power to compete^[11]. Together, these forces have normalized unprecedented systems of surveillance and centralized control over information flow.

• Legally, the speed of recent advances in AI have overwhelmed core rights of many democratic societies, leaving critical choices in the hands of restricted groups of engineers from similar social backgrounds. Intellectual property law and other protections of creative activity have been largely obviated by the capacity of large AI models to “remix and replace” content; privacy regimes have failed to keep up with the explosive spread of information; discrimination law is woefully unsuited to address issues raised by the potential emergent biases of black box AI systems. The engineers who could potentially address these issues, on the other hand, typically work for profit-seeking companies or the defense sector, come overwhelmingly from a very specific educational and demographic background (typically white or Asian, male, atheist, highly educated, etc.). This has challenged the core tenets of democratic legal regimes that aim to represent the will of the broad society they govern^[12].

• Economically, there is growing evidence that AI and the related broader tendency of information technology since the mid-1980s to replace rather than complement (especially low-educated) human labor has been a central factor in the dramatic rise in the share of income accruing to capital (rather than labor) in past decades and thereby has been a core cause of increased income inequality in developed countries.^[13] A rise in market power, mark-ups and (less consistently) industrial concentration around the world has accompanied this decline in labor’s share, particularly in countries and sectors that have most heavily adopted information technology^[14].

• (Geo-)Politically, the above forces have strengthened authoritarian regimes and political movements against democratic countries. Creating both the tools and incentives for mass surveillance, AI, and other large-scale data processing tools, has made it easier for governments to directly maintain censorship and social control. Indirectly, by concentrating economic power and the levers of social control in a small set of (often corporate) choke points, the increase in capital income and market power and the increasing authority of small groups of engineers have made it easier for authoritarian regimes to manipulate or seize the “commanding heights” of the economy and society when they wish^[15].

Furthermore, these two threats intersect; authoritarian regimes have increasingly harnessed the “chaos” of social media and cryptocurrencies to sow internal division and conflict in democratic countries. Centralized social media platforms have leveraged AI to optimize user engagement with their services, often helping to fuel the centrifugal tendencies of misinformation and opinion clustering. Yet, even when they are not actively complementing each other and may in many ways have opposite motivations, both forces have pressured democratic societies and helped undermine confidence in them, confidence that is now at its lowest ebb in much of the developed democratic world since it has been measured.^[16]

Democracies’ hostility to technology

Yet the hostilities have been far from one-sided. Democracies have, by and large, returned this hostility, viewing technology increasingly as a monolithic. Where once the public sector in democratic countries was the global driving force behind the development of information technology (e.g. the first computers, the internet, global positioning satellites), today most democratic governments are focused instead on constraining its development and are failing to respond to both opportunities and challenges it creates.

This failure has manifested in four ways. First, public opinion in democratic countries and their policy-makers are increasingly hostile to large technology companies and even many technologists, a trend commonly called the “techlash”. Second, democratic countries have significantly reduced their direct investment in the development of information technology. Third, democratic countries have been slow to adopt technology in public sector applications or that require significant public sector participation. Finally, and relatedly, democratic governments have largely failed to address the areas where most technologists believe public participation, regulation, and support are critical to technology advancing in a sustainable way, focusing instead on more familiar social and political problems^[17].

Public and policymaker attitudes towards technology took a decidedly negative turn during the 2010s. While at the end of the 2000s and early 2010s, social media and the internet were seen as forces for openness and participation, in the late 2010s they were widely blamed in commentary and to a lesser extent in public opinion surveys for many of the ills listed above ^[19]. This shift in attitudes has perhaps been most clearly reflected in elite attitudes, with best-selling books on technology, such as Weapons of Math Destruction by Cathy O’Neil and The Age of Surveillance Capitalism by Shoshanna Zuboff and films like The Social Dilemma, dominating the public conversation and political leaders across the spectrum (e.g., Jeremy Corbyn on the left and Josh Hawley on the right) taking an increasingly pessimistic and aggressive tone on the technology industry. The techlash rose to prominence to describe these concerns and is pictured in Figure A. This has been reinforced by the rise of a “cancel culture” that often harnesses social media to attack or reduce the cultural currency of prominent figures and has frequently targeted leaders in the technology industry.

Regulators in both the EU and US have responded with a range of actions, including dramatically increased antitrust scrutiny of leading technology companies, a series of regulatory interventions in Europe including the General Data Protection Regulation and the trio of the Data Governance Act, the Digital Markets Act and the Digital Services Act. All these actions have clear policy rationales and could well be part of a positive technology agenda. However, the combination of negative tone, relative disconnection from naturally allied developments in technology and general reticence on the part of commentators and policy-makers in developed democracies to articulate a positive technology vision has created an impression of an industry under siege.

Perhaps the clearest quantitative mark for this declining proactive public interest in information technology has been falling public expenditures on research and development (R&D) as a share of gross domestic product (GDP), especially in information technology. In the great majority of developed democracies, public sector research and development expenditure as a share of GDP has been declining in recent decades even as business spending on R&D has dramatically expanded and spending by the PRC (People's Republic of China) government has dramatically increased as a share of GDP and focused on information technology.^[20] Figure B shows the example of the US.

Beyond this quantitative story, the declining appearance of public financial support for information technology development has been at least as dramatic. Where once the public sector took the lead in developing what became the internet (in the US), foundations of the personal computer and analogous projects in other democratic countries (e.g., France’s Minitel), today almost all major breakthroughs in information technology are driven by the private sector.^[22]

While the original internet was almost entirely developed by the public and academic sectors (see Chapter "The Lost Dao" below) and based on open standards, the "Web 2.0" wave that dominated the late first two decades of the new millennium and the recent movements around “web3” and decentralized social technologies have received virtually no public financial support, as governments in democratic countries struggle to explore the potential of digital currencies, payments, and identity systems. While many of the most fundamental advances in computing arose from democratic governments during World War II and the Cold War, today governments have played virtually no role in the breakthroughs in “foundation models” that are revolutionizing computer science. In fact, OpenAI Founders Sam Altman and Elon Musk report having initially sought government funding and only having turned to private, profit-driven sources after being repeatedly turned down; OpenAI went on to develop the Generative Pretrained Transformer (GPT) models that have increasingly captured the public’s imagination about the potential of AI.^[23] Again, this contrasts sharply with authoritarian regimes, like the PRC and the United Arab Emirates, that have laid out and to a large extent successfully deployed ambitious public information technology strategies, including developing their own cutting-edge competitors to GPTs.^[24]

This lack of public sector engagement with technology extends beyond research and development to deployment, adoption, and facilitation. The easiest areas to measure this are the quality and availability of digital connectivity and education. Here the data are somewhat mixed, as many high-functioning democracies (such as the Scandinavian countries) have high quality and high availability internet. But it is striking that leading authoritarian regimes dramatically outperform democracies at similar development levels, especially in the latest connectivity technology. For example, according to Speedtest.net, the PRC ranks 16th in internet speeds in the world, while only 72nd in income per head; Saudi Arabia and other Gulf monarchies similarly punch above their weight^[25]. Performance on 5G, the latest generation of mobile connectivity, is more dramatic: a range of surveys find Saudi Arabia and the PRC consistently in the top 10 best-covered jurisdictions by 5G, far above their income levels.

More central to the heart of governmental responsibility in democracies, however, is the digitization of public services. Many middle-income and wealthy democracies invest less in e-government compared to authoritarian counterparts. The UN e-government development index (EGDI) is a composite measure of three important dimensions of e-government, namely: provision of online services, telecommunication connectivity, and human capital. In 2022, several authoritarian governments ranked highly, including UAE (13th), Kazakhstan (28th), and Saudi Arabia (31st), ahead of many democracies including notably, Canada (32nd), Italy (37th), Brazil (49th), and Mexico (62nd).^[26]

Digitization of conventional public services is perhaps the least ambitious dimension along which one might expect democracies to advance in adopting technology. Technology has redefined what services are relevant and in these novel areas, democratic governments have almost entirely failed to keep up with changing times. Where once government-provided postal services and public libraries were the backbone of democratic communication and knowledge circulation, today most communication flows through social media and search engines. Where once most public gatherings took place in parks and literal public squares, today it is almost a cliché that the public square has moved online. Yet democratic countries have almost entirely ignored the need to provide and support digital public services. While privately-owned Twitter is the target of constant abuse by public figures, its most important competitor, the non-profits Mastodon and the open Activity Pub standard on which it runs have received a paltry few hundreds of thousands of dollars in public support, running instead on Patreon donations.^[27] More broadly, open source software and other commons-based public goods like Wikipedia have become critical public resources in the digital age; yet governments have consistently failed to support them and have even discriminated against them relative to other charities (for example, open source software providers generally cannot be tax-exempt charities). While authoritarian regimes plow ahead with plans for Central Bank Digital Currencies, most democratic countries are only beginning explorations.

Most ambitiously, democracies could, as so many autocracies have been doing, help facilitate radical experiments with how technologies could reshape social structures. Yet, here again, democracy seems so often to stand in the way rather than facilitate such experimentation. The PRC government has built cities and reimagined regulations to facilitate driverless cars, such as Shenzhen, and has more broadly built a detailed national technology strategy covering nearly every aspect of policy, regulation and investment.^[28] Saudi Arabia is busy building a new smart city in the desert, Neom, to showcase a range of green and smart city technology, while even the most modest localized projects in democratic countries, such as Google’s Sidewalk Labs, have been swamped by local opposition.^[29]

Even when it comes to areas where technologists agree regulation and caution are critical, democracies are falling ever further behind the needs of the industry to find solutions to social challenges. There is a growing consensus among technologists that a range of emerging technologies may pose catastrophic or even existential risks that will be hard to prevent after they start to emerge. Examples include artificial intelligence systems that could rapidly self-improve their capacities, cryptocurrencies that could pose systemic financial risks, and the development of highly contagious bioweapons. They regularly bemoan the failure of democratic governments to even contemplate much less plan to confront such risks. Yet, beyond these catastrophic possibilities, a whole range of new technologies require regulatory change to be sustainable. Labor law misfits geographically and temporally flexible work empowered by technology. Copyright is far too rigid to deal with the attribution of value to data inputs to large AI models. Blockchains are empowering new forms of corporate governance that securities laws struggle to make sense of and are often put into legal jeopardy.

Yet while bold experiments with new visions of the public sector are more common in autocracies, there is an element far more fundamental to democracy itself: the mechanisms of public consent, participation, and legitimation, including voting, petitioning, soliciting of citizen feedback and so forth. Voting in nearly all democracies occurs for major offices once every several years according to rules and technologies that have been largely unchanged for a century. While citizens communicate instantaneously across the planet, they are represented in largely fixed geographic configurations at great expense with low fidelity. Few modern tools of communication or data analysis are regular parts of the democratic lives of citizens.

At the same time, autocracies have increasingly harnessed the latest digital innovations to empower their regimes of surveillance (for good and ill) and social control. For example, the PRC government has widely used facial recognition to monitor population movements, has encouraged the adoption of its Digital Yuan and other surveilled digital payments (while cracking down on more private alternatives) to facilitate financial surveillance, and has even worked on developing a comprehensive “social credit score” that would track a wide range of citizen activities and condense them to a single and widely-consequential “rating”.^[30] For several years, the Russian government has been using facial recognition to determine who is participating in protests and detain them after the fact, allowing it to remove dissenters on a large scale with much lower risks to the regime or its police forces.^[31] These techniques have intensified and have also been used to enforce war conscription since the full-scale invasion of Ukraine in February 2022.^[32] In some sense, democracy is being left behind by technology as much by its neglect of technology, compared to many authoritarian states' eager willingness to embrace it for their own ends, as by any anti-democratic tendencies of technology itself.

You get what you pay for

How did we end up here? Are these conflicts the natural course of technology and of democratic societies? Is a different future possible?

A range of work suggests that technology and democracy could co-evolve in a diversity of ways and that the path most democracies are on is a result of collective choices they have made through policies, attitudes, expectations, and culture. The range of possibilities can be seen through a variety of lenses, from science fiction to real-world cases.

Science fiction shows the astonishing range of futures the human mind is capable of imagining. In many cases, these imaginings are the foundation of many of the technologies that researchers and entrepreneurs end up developing. Some of these correspond to the directions we have seen technology take recently. In his 1992 classic, Snow Crash, Neal Stephenson imagines a future where most people have retreated to live much of their lives in an immersive “metaverse”.^[33] In the process they undermine the engagement necessary to support real-world communities, governments, and the like, making space for mafias and cult leaders to rule and develop weapons of mass destruction. This future closely corresponds to elements of the “antisocial” threats to democracy from technology we discussed above. Stephenson and other writers further extend these possibilities, which have had a profound effect in shaping technology development; for example, Meta Platforms is named after Stephenson’s metaverse. Similar examples are possible for the tendency of technology to concentrate power through creating “superintelligences” as in the fiction of Isaac Asimov and Ian Banks, the predictive futurism of Ray Kurzweil and Nicholas Bostrom, and films like Terminator and Her.^[34]

But these possibilities are both very different from each other and are far from the only visions of the technological future to be found in sci-fi. In fact, some of the most prominent science fiction shows very different possibilities. Two of the most popular sci-fi television shows of all time, The Jetsons and Star Trek, show futures where, respectively, technology has largely reinforced the culture and institutions of 1950s America and one where it has enabled a post-capitalist world of diverse intersecting alien intelligences (on which more below). But these are two among thousands of examples, from the post-gender and post-state imagination of Ursula Le Guin to the post-colonial futurism of Octavia Butler. All suggest a dizzying range of ways technology could coevolve with society^[35].

But science fiction writers are not alone. The primary theme of the field of Science and Technology Studies, including the philosophy, sociology, and history of science, has been the contingency and possibility inherent in the development of science and technology and the lack of any single necessary direction for their evolution^[36]. These conclusions have been increasingly accepted in social sciences, like political science and economics, that traditionally viewed technology progress as fixed and given. Two of the world’s leading economists, Daron Acemoglu and Simon Johnson, have recently published a book that argues that the direction of technological progress is a key target for social policy and reform while documenting the historical contingencies that led to the directions of technology we have seen in the past^[37].

Perhaps the most striking illustration comes from comparing the ways technology has advanced across countries today. Where once leading thinkers predicted the power of technology to sweep away social differences, today the technological systems of powers great and sometimes small define their competing social systems as much as their formally stated ideologies: the PRC surveillance regime looks like one technological future, while the Russian hacking networks seem another, the growing space of web3-driven communities a third, the mainstream Western capitalist countries on which we have focused a fourth and the heterogeneous digital democracies of India, Estonia, and Taiwan something else entirely that we will explore in depth below. Also possible, largely along the web3 community-driven approach, is an African model that could be built on open source and interoperability, reflective of the communal inclination of many African cultures. Far from converging, technology seems to be proliferating possible futures.

So, if our current trajectory of technology and social relationship to it in Western liberal democracies is not inevitable, in what ways are we choosing to be on this conflictual path? And how might we get off it?

While there are many ways to describe the choices democratic societies have made about technology, perhaps the most concrete and easiest to quantify are the investments realized. These show clear choices about technological paths that Western liberal democracies (and thus most of the financial capital in the world) have made about investments in the future of technology, many are of quite recent origin. While these have recently been driven primarily by the private sector, they reflect earlier priorities set by governments that are in many ways just beginning to filter through to private sector applications.

Beginning with recent trends in the increasingly well-measured venture capital industry, the last decade has seen a dramatic and overwhelming focus of venture capital within the high technology sector into artificial intelligence and cryptocurrency-adjacent “web3” technologies. Figure C displays data on private investment in AI collected by NetBase Quid and charted by Stanford’s Center for Human-Centered Artificial Intelligence’s 2022 AI Index Report, showing its explosive growth over the course of the 2010s, growth that has come to dominate private technology investment; Figure D shows the same (over a different period and quarterly) for the web3 space based on data from Pitchbook.

However, while these priorities are relatively recent and appear to emerge from the logic of “the market”, they reflect a much longer-running and collectively direct set of choices. These stem from the investments governments in democratic countries have made. ^[40].

Furthermore, these investments are not just choices that could have been made differently; they are quite recent and were made very differently immediately prior. These investments are reflected in the canonical technologies of the last few decades. Artificial intelligence was heralded as a coming revolution throughout much of the 1980s, as reflected in Figure E showing the relative frequency of this phrase in English books as tracked by Google nGrams. Yet the defining technology of the 1980s was quite opposite: the personal computer that made computing a complement to individual human creativity. The 1990s were haunted by Stephenson’s science fictional imagination of the possibilities of escapist virtual worlds and atomizing cryptography, the connective tissue of the internet swept the world, ushering in an unprecedented age of communication and cooperation. Mobile telephony in the 2000s, social networking in the 2010s, and the scaffolding of remote work in the 2020s…none of these have focused on either cryptographic hypercapitalism or artificial superintelligence. This reflects, with an extensive lag, the shift in investments made by public sector research funders away from supporting these technologies and towards investment in cryptography and artificial intelligence, as we discuss and document in "The Lost Dao" below, driven by a variety of (geo)political factors.

Ideologies of the twenty-first century

If the path of technology is not predetermined and instead can be significantly shaped by collective choices regarding investment, how should we think about the flexibility we have as a society in choosing among possible directions? How much scope is there for choice and what do these look like?

One useful analogy for thinking about choice over directions a society might take is ideologies. It is "common sense" that different societies have chosen or might choose to organize themselves in terms of different (combinations of) ideologies: communism, capitalism, democracy, fascism, theocracy, etc. Each of these incorporates strengths and weaknesses, appeals to some more than to others, and coheres and prescribes to differing degrees. There may be configurations of these ideologies that simply do not work or require specific historical and social conditions.

We might look at different trajectories for technology in a similar way. The range of futures is not unlimited or infinitely malleable: some things are easier, harder, or outright impossible. But neither is it predetermined. There are clusters of plausible visions of the future and technologies that empower these; through our collective technological investments, we help choose among these possibilities.

While a bit less familiar than the linear and progressive story about technology that is most common today, this perspective is very far from original. It is a recurring theme in literature, scholarship, and even entertainment. One striking example is the series of computer games Civilization created by Sid Meier, in which the player charts a course for a people from prehistory to the future. A defining characteristic of the game is the diversity of possible technological paths and the way these interact with social systems a society may adopt.

The latest entry in the series, Civilization VI, and specifically the “Gathering Storm” expansion pack for it, illustrates our argument quite elegantly. In that game, the “Information Era” features a choice among three ideologies: “Synthetic Technocracy,” “Corporate Libertarianism”, and “Digital Democracy”, with corresponding strengths, weaknesses, and connections to technological development. While the names for each of these are a bit awkward and we shorten them below, we will argue in what follows that they do a good job describing in broad strokes, like Communism, Fascism and Democracy in the 20th century, the great techno-ideological debate of our time.

Artificial Intelligence and technocracy

The first and most widely expressed vision of the future of technology today centers around Artificial Intelligence (AI) and the way social systems will have to adapt to it; it is captured by the Civilization VI “Synthetic Technocracy” category, or Technocracy for short.

Technocracy focuses on the potential of AI to create what OpenAI Founder Sam Altman calls “Moore’s Law for Everything”: a transformation where AI makes all material goods cheap and abundant and thus allows the abolition, at least in principle, of material scarcity.^[42] Yet this potential abundance may not be equally distributed; it is plausible that its value will concentrate in a small group that controls and directs AI systems. A key element of the technocratic social vision is therefore material redistribution, usually through a “universal basic income” (UBI). Another central focus is on the risk of AI(s) getting out of human control and threatening human survival, and thus on the need for strong and often centralized control over who has access to these technologies, as well as ensuring they are built to faithfully execute human desires. While the precise contours differ across the exponents of this view, the idea of “Artificial General Intelligence” (AGI) is central: machines that exceed human capabilities in some generalized way, leaving little measurable utility in human individual or collective cognition.

Leading exponents of this view in Silicon Valley are Altman and his mentor Reid Hoffman and until recently Altman’s OpenAI co-founder Elon Musk. The view is also popular in the PRC, where it has been advanced by Jack Ma, economist Yu Yong-Ding and even by the PRC’s official New Generation Artificial Intelligence Development Plan with a strong reliance on the Marxist idea of “central planning”. It also appears throughout science fiction, particularly the work of authors mentioned above like Asimov, Banks, Kurzweil and Bostrom. Bostrom's latest book, Deep Utopia: Life and Meaning in a Solved World, is perhaps the purest expression of this view.^[43] Leading organizations aligned with this perspective include OpenAI, DeepMind, and other advanced artificial intelligence projects. The political campaigns of Andrew Yang in the United States helped bring this perspective to the mainstream of politics and technocratic ideas show up in toned-down forms in much of the thought of the “tech left”, including commentators like Ezra Klein, Matthew Yglesias, and Noah Smith.

Crypto and hyper-capitalism

A second view is much less common in the mainstream media, but has been a dominant theme in the community that has built around Bitcoin and other cryptocurrencies and in various related internet communities; it is captured by the Civilization VI “Corporate Libertarianism” category, which we will abbreviate to "Libertarianism" below.

Libertarianism focuses on the potential (or in some telling inevitability) of cryptography and networking protocols supplanting the role of human collective organization and politics, liberating individuals to participate in unfettered markets free from government and other collective “coercion” and regulation. Fiction has been the central inspiration for libertarian thinking, including the work of Ayn Rand and Stephenson.^[44] Stephenson’s books, especially Snow Crash discussed above and Cryptonomicon (1999), while seemingly and expressly intended as dystopian warnings, have been adopted as blueprints by adherents of Libertarianism.^[45] Exemplary technologies in these works and have since become central to the Libertarian community are immersive virtual worlds (viz. Stephenson's metaverse), digital currencies independent of governments, private sovereignties especially based in ungoverned spaces such as floating cities or “seasteads” and strong cryptography as a means of evading collective control/law. The Bitcoin, Web3, 4Chan, and other “peripheral” but influential online communities are core to the social base of the Libertarian perspective.

Perhaps partly because it is less mainstream than Technocracy, libertarianism has a much clearer intellectual canon and set of leaders. The Sovereign Individual by James Dale Davidson and Lord William Rees-Mogg, the writings of Curtis Yarvin under the pen name Mencius Moldbug, The Network State by Balaji Srinavasan and Bronze Age Mindset are widely read and cited in the community.^[46] Venture capitalist Peter Thiel is widely seen as the central intellectual leader, along with others (such as the authors mentioned) whom he has funded or promoted the work of.

Libertarianism has a close, but also somewhat complicated, relationship with the nationalist and far-right in democratic countries. On the one hand, most participants identify with this group and support it, to the extent they engage with politics, as illustrated by Thiel’s emergence as a primary financial supporter of Donald Trump and his supporters. In fact, several leading hard right politicians are closely connected to the Libertarian worldview: prominent British Conservative Member of Parliament Jacob Rees-Mogg is the son of Lord William Rees-Mogg, Thiel employs former Austrian Chancellor Sebastian Kurz and Thiel protégés Blake Masters and J. D. Vance ran for Senate in 2022, with the latter winning a seat.

On the other hand, Libertarianism is consistently hostile to nationalism (or any other form of collectivism or solidarity) and Libertarian followers routinely mock and dismiss many core religious, national and cultural values associated with the right. This apparent contradiction may be resolved by a shared antipathy to what they perceive as dominant left-wing cultural values or by an “accelerationist” attitude as advocated by Yarvin, Davidson and Rees-Mogg that views the “nationalist backlash” to the inevitable technological trends as an accelerant and possible ally in the dissolution of the nation-state.

Stagnation and inequality

These two ideologies have, to a significant extent though often in moderated form, dominated public imagination about the future of technology in most liberal democracies and thus shaped the direction of technology investment for most of the last half-century. While the Technocratic story sounds fresh and related to recent progress in AI, related discussions around AI were almost as fever pitched as far back as the 1980s, as illustrated by Figure E. While the recent discussions around web3 technologies have raised its profile, Libertarianism was arguably at its peak in the 1990s, with John Perry Barlow’s “Declaration of the Independence of Cyberspace”, Stephenson’s novels and the publication of The Sovereign Individual.

The radical promises of these visions led many to anticipate dramatic economic and productivity growth from information technology, as well as the waves of privatization, deregulation and tax cuts that went along with them in most liberal democracies beginning roughly half a century ago. Yet these promises are far from bearing fruit and economic analysis increasingly suggests these directions for technology may play a key role in explaining that failure.

Instead of the promised explosion of economic possibility, the last half-century has seen a dramatic deceleration of economic and especially productivity growth. Figure F shows the growth in the United States of “Total Factor Productivity (TFP)”, economists’ most inclusive measure of the improvement in technology, averaged by decades from the beginning of the 20th century to today. Rates during the mid-century “Golden Age” roughly doubled their levels both before and after during the period we dub the “Digital Stagnation”. The pattern is even more dramatic in other liberal democratic countries in Europe and in most of democratic Asia, with South Korea and Taiwan notable exceptions.

To make matters worse, this period of stagnation has also been one of dramatically rising inequality, especially in the United States. Figure G shows average income growth in the US by income percentile during the Golden Age and Great Stagnation respectively. During the Golden Age, income growth was roughly constant across the distribution, but trailed off for top-income earners. During the Digital Stagnation, income growth was higher for higher earners and only exceeded the average level during the Golden Age for those in the top 1%, with even smaller groups earning the great majority of the overall much lower income gains.

What has gone so wrong in the last half-century compared to the one before? Economists have studied a range of factors, from the rise of market power and the decline of unions to the progressively greater challenge of innovating when so much has already been invented. But increasing evidence focuses on two factors closely tied to the influence of technocracy and libertarianism respectively: the shift in the direction of technological progress towards automation and away from labor augmentation and the shift in the direction of policy away from proactively shaping industrial development and relations and towards an assumption that “free markets know best”.

On the first point, in a series of recent papers, Acemoglu, Pascual Restrepo, and collaborators have documented the shift in the direction of technical progress from the Golden Age to the Digital Stagnation. Figure H summarizes their results, plotting cumulative changes in productivity over time from labor automation (what they call “displacement”) and labor augmentation (what they call “reinstatement”)^[49]. During the Golden Age, reinstatement roughly balanced displacement, leaving the share of income going to workers essentially constant. During the Digital Stagnation, however, displacement has slightly accelerated while reinstatement has dramatically fallen, leading to slower overall productivity growth and a significant reduction in the share of income going to workers. Furthermore, their analysis shows that the inegalitarian effects of this imbalance have been exacerbated by the concentration of displacement among low-skilled workers.

The role of “neoliberal” policies in contributing to the stagnation and inequality of this period is widely debated and we suspect most readers have formed their own views on the matter. One of us was also co-author of a book that contains a review of the evidence as of roughly a decade ago.^[51] We will thus not go into detail here and refer readers to instead to that or other related writing.^[52] However, clearly, the defining ideological and policy direction of this period was an embrace of capitalist market economics, often closely tied to claims that such an embrace was necessitated by the globalization of technology and the resulting impossibility of collective governance/action that is core to the Libertarian ideology. The, largely failed, last half-century of technology and policy has thus been characterized by the dominance of Technocracy in the sphere of technology and Libertarianism in the sphere of policy.

Of course, the last half century has hardly been devoid of technological breakthroughs that have genuinely brought about positive, if uneven and sometimes fraught, transformations. Personal computers empowered unprecedented human creativity in the 1980s; the internet allowed communication and connection across previously unimaginable distances in the 1990s; smartphones integrated these two revolutions and made them ubiquitous in the 2000s. Yet, it is striking that none of these most canonical innovations of our time fit neatly into the Technocratic or Libertarian stories. They were clearly all technologies that augmented human creativity, often called “intelligence augmentation” or IA, rather than AI.^[53] Yet neither were they envisioned primarily as tools to escape existing social institutions; they facilitated rich communication and connection rather than market transactions, private property, and secrecy. As we will see, these technologies emerged from a very different tradition than either of these two. Thus, even the few major technological leaps in this period were largely independent of or in contrast to these visions.

A fraying social contract

Yet the economic conditions surrounding the embrace of Technocracy and Libertarianism are only the easiest to quantify and thus most headline-grabbing. Deeper, more insidious, and ultimately more damaging have been the corrosion of the confidence, faith, and trust on which social support of both democracy and technology rest.

Faith in democratic institutions has been falling, especially in the last decade and a half in all democracies, but especially in the US and developing democracies. In the US, dissatisfaction with democracy has gone from being the opinion of a fringe (less than 25%) to being the majority opinion in the last 3 decades.^[54] While it is less consistently measured, faith in technology, especially leading technology companies, has been similarly declining. In the US, the technology sector has fallen from being considered the most trusted sector in the economy in the early and mid 2010s to amongst the least trusted, based on surveys by organizations like the Public Affairs Council, Morning Consult, Pew Research and Edelman Trust Barometer.^[55]

These concerns have spilled out more broadly to a general loss of faith in a range of social institutions. The fraction of Americans expressing high confidence in several leading institutions (including organized religions, federal governments, public schools, media, and law enforcement) has fallen to roughly half its level when such surveys began, around the end of the Golden Age in most cases.^[56] Trends in Europe are more moderate and the global picture is uneven, but the general trend towards declining institutional confidence in democratic countries is widely accepted.^[57]

Reclaiming our future

Technology and democracy are trapped between two sides of a widening gulf. That war is damaging both sides of the conflict, undermining democracy and slowing technological development. As collateral damage, it is slowing economic growth, undermining confidence in social institutions, and fueling inequality. This conflict is not inevitable; it is the product of the technological directions liberal democracies have collectively chosen to invest in, once fueled by ideologies about the future that are antithetical to democratic ideals. Because political systems depend on technologies to thrive, democracy cannot thrive if we continue down this path.

Another path is possible. Technology and democracy can be each other’s greatest allies. In fact, as we will argue, large-scale “Digital Democracy” is a dream we have only begun to imagine, one that requires unprecedented technology to have any chance of being realized. By reimagining our future, shifting public investments, research agendas, and private development, we can build that future. In the rest of this book, we hope to show you how. And we will begin by telling you the story of a place that has gone farther than any other in realizing that future, a place where democracy and digital technology are not just allies, but deeply mutually entwined.