If there is one thing that seems to be uniting many policymakers, pundits, and scholars across the political spectrum, it is that the internet and digital media are not the inevitably liberating forces they were purported to be. Apple’s famous 1984 Super Bowl commercial cast its new Macintosh personal computer as a renegade force against an Orwellian Big Brother. But today private companies have accomplished what the twentieth-century state could only dream of: comprehensive tracking of the movements, communications, and purchases of individuals in the United States and across the world.

In her recent essay for Hypertext, Jennifer Burns argues, “Digital media is acting on our politics in obvious ways, but our politics don’t seem to be acting much on digital media.” However, the problems we face today are, unfortunately, not simply the result of political inaction. Instead, they reflect failures of past policymaking. Legal regimes designed long ago have left us with a system that continues to favor the interests of major companies while doing little for many Americans’ concerns about digital media. Of course, we will need to respond more effectively to recent innovations in digital technology. However, we must ensure reforms target the long-standing incentives shaping the tech giants’ decision-making — and recognize that the broader political context matters, too. Often, the features of online life we find most objectionable are symptoms of two deeper problems: (1) the industry’s reliance on digital advertising, which drives its aggressive collection of user data and (2) broader political dynamics that are spurring affective polarization, extremism, and the “dark passions” of hatred and resentment. This essay will focus on (1), as fixing (2) requires a project of liberal-democratic revitalization no digital policy reform alone will be able to accomplish.

Developing an effective tech policy agenda for the 21st century demands conceptual clarity on two dimensions. First, what are the kinds of problems that we are most worried about? To what degree are these problems rooted in technological innovation itself or in the interaction of technology with our political-economic structures? Second, how have policymakers already attempted to govern digital technologies, and what can we learn from these past attempts? It is imperative we keep these questions top of mind so that we avoid wasting valuable political energy on legislative efforts that might sound nice but only exacerbate the very problems they purport to address.

Diagnosing the pathologies of the digital revolution

The most urgent concerns about digital technologies fall under the following categories:

• Fairness, discrimination, and algorithmic bias.

• Disinformation, extremism, and polarization.

• Harmful content, including harassment, abuse, and hate crimes.

• “Addictive” properties of smartphones, social media platforms, and online gambling.

• Free speech and censorship.

• Intellectual property, copyright protections, and rights to publicity and/or personality.

• Data governance, including privacy and portability of data.

• Digital surveillance by public and private authorities.

Many of these problems are cast as failures of content moderation, but that diagnosis just scratches the surface. The deterioration of discourse on the modern internet, including popular social media platforms, is a symptom of a deeper, structural condition: the reliance of most companies on a digital advertising business model. This business model means that companies are incentivized to engage in pervasive data collection and analysis — also known as surveillance —to track and maximize user engagement, and thus advertisers’ click-through rates. That imperative drives product and design choices from top to bottom.

It is the power of surveillance-based targeting that makes many of the aforementioned concerns, like online extremism, so worrisome. Noxious content can be pushed to the people most susceptible to its message because of the data that companies collect about them. Policies that focus on moderating content risk censoring speech without fixing the underlying data governance regime that encourages companies to push that speech into the corners where it will do the most damage.

Nor will better moderation of online content solve the larger cultural problems of which it is as much symptom as cause. Members of traditional media, such as cable news, are also responsible for the mainstreaming of extremism and disinformation.1 And, of course, we cannot forget the public commentators and politicians who legitimate and perpetuate disinformation and violent rhetoric and policies. By failing to hold accountable leaders who take illegal, illiberal, and anti-democratic actions, whether through social censure or appropriate legal and political methods policymakers have forsaken powerful remedies for refuting disinformation and extremism while rebuilding trust within the populace. Political extremism online has become incorporated into this broader fabric of partisanship.

Reforms that target online content moderation are unlikely to resolve problems driven by deeper structural issues with the business model of the modern internet or the pathologies of our politics. In fact, they can backfire by making it harder for vulnerable communities to organize while permitting the giants to comply with the letter of the law but exploit loopholes.

Consider, for example, a recent example of American politicians acting on digital technology companies: the 2018 Allow States and Victims to Fight Online Sex Trafficking Act, known as FOSTA-SESTA because it incorporated elements of another bill known as the Stop Enabling Sex Traffickers Act.

Learning from past political decisions to govern digital media

FOSTA-SESTA amended Section 230 of the 1996 Communications Decency Act, which attempted to reduce pornographic online content. While the Supreme Court struck down most of the Communications Decency Act for violating the First Amendment’s speech protections, Section 230 survived. This section protected online service providers, and now social media platforms, from liability for content users shared. In large part, it responded to the uncertainty produced by two major court decisions regarding early internet forums. Before Section 230, had they moderated comments on early discussion boards, online service providers would be treated as publishers, not distributors, of that content. Consequently, they would be liable for violations of defamation and obscenity laws. This created perverse incentives to either let user-generated content run wild or simply stop hosting such material entirely. In short, Section 230 enables online service providers to moderate content at all. Section 230 does not excuse companies from a plethora of responsibilities, such as intellectual property laws and the enforcement of federal criminal laws. But it enables them to exercise their speech rights to moderate content and host user-driven sites that billions of people find useful.

FOSTA-SESTA removed Section 230’s liability protections with regard to content that promoted sex-trafficking. However laudable its intentions, the law seems to have backfired. Since the adoption of FOSTA-SESTA in and unrelated seizure of backpage.com in April 2018, authorities have found it harder to collect evidence of sex trafficking as platforms have moved overseas and many people involved in the trade now rely on encrypted and disappearing messages on social media platforms.

Victims of sex traffickers and other sex workers must also be able to communicate online and warn each other about dangerous actors. Such protective practices might also seem like “surveillance” or monitoring, but they are survival strategies embraced by a relatively less powerful population. Scholars use the term “sousveillance” to denote the difference. A 2020 survey of sex workers’ experiences in the wake of the 2018 changes found that FOSTA-SESTA gave them “a general sense fear and paranoia” about what information they could legally and safely share online, making it harder to rely on the internet as a tool for finding community, support, and resources for verifying the trustworthiness and safety of new clients.2 While a 2023 D.C. Circuit ruling helped affirm that “sex workers, advocates for sex-workers’ rights, and other online speakers were protected from prosecution,” ambiguities about the law and thus its potential chilling effect remain.

The example of FOSTA/SESTA suggests the potential limits of policy responses that prioritize reducing the presence of morally objectionable content online. Such an approach might successfully roll back undesirable content on mainstream, market-dominant, and domestic websites. But it may accomplish this while also reducing legal, less objectionable, or even morally desirable content, including that which would be produced by the very people that the policy is supposed to serve. It makes it harder for public and private actors to find reliable information about dangerous individuals and groups.

I do not believe taking seriously the criticisms of FOSTA-SESTA means we must give up on the possibility of policy solutions to problems we see online. Rather, within this story we can make out the contours of an alternative paradigm that understands digital life to be part and parcel of our broader political economy. While seeking to hold companies responsible for social harms, we must not lose sight of how ordinary individuals use the internet to empower themselves and their communities, and whether private and public policies undermine or support these capacities.

The internet’s liberating potential is undermined less by objectionable content and more by the massive data collection, analysis, and use that render those individuals and anyone to whom they are connected vulnerable to private and public forms of surveillance. Online data is fundamentally relational, so that companies can collect data about non-users. That means data you provide to private companies can be used to police and punish other people. Consider, for example, users of a popular Muslim prayer app, Muslim Pro, who felt exploited and complicit in the oppression of their peers when they found out the company’s data had wound up being shared with the American military. It is this fundamental business of data that shapes most of the choices that technology companies make, including the promotion of noxious or seemingly “addictive” content. We are likely to see these dynamics of so-called “surveillance capitalism” continue to play out as chatbot companies search for profit models, as seen in OpenAI’s turn to digital advertising.

Privacy reforms need not pursue a total ban on digital advertising; the European Union’s General Data Protection Regulation has shifted companies’ incentives away from more personally invasive behavioral targeting to contextual targeting, a trend that American privacy policy could further encourage.3 Federal privacy legislation could limit companies to collecting only the data required to offer an effective service for their consumers; empower users by making further data collection opt-in rather than the default; strengthen existing protections against discriminatory algorithmic targeting with respect to housing, employment, and banking; and prohibit state actors and police from circumventing Fourth Amendment protections against warrantless search and seizures by purchasing data from brokers. To be clear: Because challenging the status-quo data governance regime that favors companies over users would harm the bottom line of most major digital technology companies, it is an unavoidably ambitious project. But if we think we are at a revolutionary moment that will define the future of our ever-more digital political economy, then the time for ambition is now.

Kristen Collins is a Senior Fellow with the F.A. Hayek Program for Advanced Study in Philosophy, Politics, and Economics and a Senior Research Fellow at the Mercatus Center at George Mason University. Her Substack is Theory of Virtual Sentiments.

Share

Last week, the Department of Defense gave Anthropic an ultimatum: Let the Pentagon use its AI models for “any lawful purpose,” without restriction, by 5:01 p.m. Friday or face the consequences. Those consequences turned out to include a “supply chain risk” designation normally reserved for foreign adversaries and firms associated with them, a presidential directive for every federal agency to cease using Anthropic’s technology, and threats to compel cooperation under the extraordinary powers of the Defense Production Act.

Anthropic refused to compromise on two red lines that ought to put us beyond the realm of “lawful purpose” anyway: no mass domestic surveillance and no fully autonomous weapons without a human in the loop. The Pentagon labeled this defiance. As commenters across the ideological spectrum have pointed out, the administration’s actions are deeply illiberal and its rationale incoherent. It’s also, of course, hard to look past the relationship that some of Anthropic’s competitors have with the administration and not wonder whether they pushed the DoD to take an excessively hard line for their own benefit. Most Americans agree with Anthropic’s position.

And yet, for all the wrong reasons, the Trump administration has tripped and fallen into an interesting question: How should the government think about preserving its own autonomy while it makes use of privately-held AI models?

The administration has attempted to answer this question with allegations that don’t really scan; suggesting, for example, that Anthropic is bad because it is a Trojan horse for effective altruism to “seize veto power over the operational decisions of the United States military” or because “there has been no bigger thief of Americans’ public identity information en masse or creators’ works than by Anthropic.”

As is so often the case, other players in the MAGA ecosystem have now been left to graft logic onto Trump and Hegseth’s impulsive behaviors with varying success. But writing on X, Acting Under Secretary of State for Foreign Assistance, Humanitarian Affairs, and Religious Freedom Jeremy Lewin made one such ex post argument that we should take seriously:

This isn’t about Anthropic or the specific conditions at issue. It’s about the broader premise that technology deeply embedded in our military must be under the exclusive control of our duly elected/appointed leaders. No private company can dictate normative terms of use — which can change and are subject to interpretation — for our most sensitive national security systems. The @DeptofWar obviously can’t trust a system a private company can switch off at any moment.

Lewin is mostly right in principle. No government could, would, or should tolerate a private vendor dictating the terms of how it fulfills its core functions. It’s anti-democratic to outsource so much of your government to private industry that you fail to translate policy goals into real results.

But here’s the thing: Even if we accept the dubious premise that this is a case in which a private company is exercising inappropriate influence over how the government executes its legitimate mission, that is not a new problem. It is, in fact, one of the most common and well-documented failures in federal operations. We just don’t usually talk about it in such dramatic terms. It manifests instead in now-expected tech meltdowns, call centers that don’t scale, and websites that are unusable to the people they need to serve. But it costs us far more in money, failed programs, and degraded public services than any harm Anthropic’s usage policy has ever caused.

And while Lewin correctly identifies this as a problem, the solution cannot be a sudden resort to punitive intervention in the market; this won’t work. Instead, it requires habitual changes inside every agency, office, and program. For anything more complex than a pencil, government contracting poses a classic principal-agent problem: It is difficult for the ostensible boss to track what the ostensible underling is actually doing. Fixing it requires doing something radical: growing the government’s own internal capacity at the expense of an industrial base that has captured large parts of the government and profits by perpetuating dysfunction.

That insight has taken decades to surface, and we struggle mightily to execute on it even now. But AI is going to pour jet fuel on this problem, because it implicates so many normative questions about humanity, ethics, and ownership. We need to ask bigger questions and remain open to bigger answers about how to address this before it’s too late.

We know how to solve vendor capture, actually

The Anthropic fight has captured the public’s interest, but most of the time, vendor capture doesn’t look like a CEO drawing ethical red lines on national television. It looks like an agency that can’t switch IT contractors because no one on the government side understands how the old system works. It looks like a contractor who has maintained a critical system for so long that all the institutional knowledge has migrated to their side of the table. It looks like contracts designed — sometimes deliberately, sometimes through sheer inertia — to make it functionally impossible for anyone else to compete for them.

The practical impact isn’t usually an ideological company imposing its values on public policy. It’s billions of dollars siphoned off through dependency; programs hobbled by vendors who face no real accountability; and a government so reliant on outside firms that it can’t perform even basic oversight of their work.

It’s easy to imagine that this is nefarious, but the reality is generally quite banal. Agencies can’t function without their contractors not because the contractors are so brilliant but because the government has let its own capacity atrophy to the point where the dependency is total. As my colleague Matthew Burton recently explained, vendor capture comes in many forms.

• There’s technological capture, where a vendor controls a government IT system so completely through proprietary code, managed-service arrangements, or contractual terms the agency failed to negotiate that the government can’t even access, modify, or migrate the system it paid to build.

• There’s intellectual capture, where all the institutional knowledge of how a critical system works has migrated to the vendor’s side of the table. This happens sometimes because the agency never required adequate documentation and sometimes because the agency’s best engineer retired and was immediately hired by the incumbent contractor at twice their government salary.

• And there’s psychological capture, where agency leaders have internalized the idea that their vendors are simply better than they are, and stop asking hard questions as a result.

In some senses, this is just a textbook example of the principal-agent problem: because vendors are profit-maximizing entities, they will behave in ways that both self-perpetuate their own incumbency and maximize the “moat” around their position, both at the expense of the very entity that hired them in the first place. This is a well-studied pathology, though one that has gotten worse in the last several decades as the complexity of 21st century technology grows.

Over the past decade, a hard-won consensus has emerged about how to address vendor capture in government. The answer isn’t to threaten vendors or nationalize their products. It’s to close what I call the sophistication gap: the gulf between a government that can barely understand and evaluate what it is buying and a vendor who seems to know everything.

This means the government needs its own technical capacity. It doesn’t need to do everything involved in deploying a system itself — that’s neither practical nor desirable. But it needs to be able to understand the process. It needs product managers who can set priorities and make tradeoffs. This involves the government hiring its own people who decide what to build and for whom, not just contracting officers who check compliance boxes. It needs its own software engineers who can review a vendor’s code, spot architectural problems, and course-correct before the project goes sideways. It needs people to maintain systems in steady state during transitions between vendors, between administrations, and over the long term. The absence of this capacity is staggeringly expensive. Many, many government IT programs fail outright, and many that did eventually launch cost billions more than expected.

We’ve seen what happens when this capacity doesn’t exist and when it does. When a several-hundred-million dollar redesign opf the Free Application for Student Financial Aid collapsed in 2023, the proximate cause was that the main vendors doing the work had produced code that simply didn’t work. But the upstream problem really was that the department ostensibly managing the vendor didn’t know the right questions to ask about their work. When Jeremy Singer, a College Board executive that the White House brought in to manage the turnaround, arrived at FAFSA he found “no ability to check the veracity of what the vendor said as far as status, quality of code — all the things you’d want to do.” Federal Student Aid wasn’t getting a good picture of the actual state of the product because it didn’t know the right questions to ask. “I don’t think the department knew how screwed they were.”

However, the Department of Education didn’t subsequently fix FAFSA by threatening its vendors with forced contracting or arbitrary risk designations. Instead, it managed to right the ship by bringing on a cadre of empowered, senior talent who were capable of reviewing whether contractors’ work was technically sound and explaining tradeoffs to policymakers. This isn’t rocket science and doesn’t require compelling private industry to act at gunpoint.

It’s telling, though, that this lesson had to be relearned. FAFSA was basically in the same failure mode as Healthcare.gov a decade earlier: well-meaning but outgunned policy professionals, unable to manage a vendor doing work they didn’t really understand. We keep making the same mistake because scaling this mindset across the largest enterprise in the country is extraordinarily hard, and because the structural incentives in federal hiring, budgeting, and procurement all push against it.

But AI is genuinely different

Here is where the DoD-Anthropic standoff points at something genuinely new, even if the administration both misunderstood its own problem and then prescribed exactly the wrong medicine.

The traditional answer to vendor capture rests on a key assumption: that if you put a sufficiently skilled person on the government’s side of the table, they can look under the hood. Traditional software is opaque but reconstructable. If a COBOL mainframe has no documentation (and many of them don’t!) you can hire a developer to trace the logic line by line and figure out what it does. It’s expensive, but the code is eventually legible to a competent reader. The architecture is knowable. The behavior is, in principle, fully explicable, even if it’s hard to recreate and expensive to maintain.

AI models are different in kind, not just in degree. The weights and parameters that determine a model’s behavior are trade secrets, but even if they weren’t, the relationship between those weights and any given output is extraordinarily difficult to trace. There is evidence of emergent behavior (i.e., capabilities that arise from training without anyone designing them in) and a long-standing focus by all frontier AI labs on preventing “misaligned” behavior. The companies building these tools do not themselves always fully understand why their models behave the way they do. In fact, AI is so illegible to its own designers that understanding it has become a major area of active research called interpretability.

Dario Amodei, the CEO of Anthropic, described this succinctly last year:

Modern generative AI systems are opaque in a way that fundamentally differs from traditional software. If an ordinary software program does something — for example, a character in a video game says a line of dialogue, or my food delivery app allows me to tip my driver — it does those things because a human specifically programmed them in. Generative AI is not like that at all. When a generative AI system does something, like summarize a financial document, we have no idea, at a specific or precise level, why it makes the choices it does — why it chooses certain words over others, or why it occasionally makes a mistake despite usually being accurate.

This difference matters enormously for government accountability. With traditional software, the strategy of closing the sophistication gap works.

With AI, it is not obvious that anyone the government hires could pop open the hood and investigate whether a model was putting its thumb on the scale — purposely or accidentally — in ways inconsistent with the public sector’s priorities. This problem is particularly dire because of the oligopolistic nature of the AI market. There are lots of engineers who could build an alternative software product with enough time and resources, but very few alternative AI models and tremendously high capital costs to train a new one. In such a highly concentrated market, one or two key players declining to work with the government is itself policy-influencing. And, while the switching costs are relatively low between models right now, the market pressure to raise them will be great.

Furthermore, the nature of AI technology is such that the ethics, values, and risk tolerances of its builders are fundamental to the product in a way that isn’t true of more traditional enterprise software. It may simply not be possible to require an AI company to check those values at the door, as the Trump administration seems to be demanding, because of that tight integration. Civil servants are acculturated to speak up but ultimately follow the president’s direction; AI models do not necessarily share this impulse.

One day, this may change as interpretability advances or the market fragments, but that day is yet a long way off. As we ask AI to perform complex tasks, this gap between what the technology can do and what the government understands becomes a significant risk to the public interest.

Imagine the shoe on the other foot

Imagine a future Democratic administration arrives at the Department of Health and Human Services and discovers that Elon Musk’s xAI has expanded its footprint beyond “generation of first drafts of documents and other communication materials” and is now deeply integrated into software looking for patterns of fraud in Medicare data. Per the terms of their contract, xAI declines to explain why the system keeps flagging supposed fraud cases overwhelmingly in blue states. The agency’s technical staff can’t look beyond the outputs they receive. They can’t audit the model’s reasoning because, in a meaningful sense, there isn’t traditional “reasoning” to audit — just statistical patterns baked into billions of parameters that no one can fully trace and that the vendor is under no obligation to provide insight into.

The incoming administrators have no choice but to cancel the contract entirely. The program grinds to a halt while they change vendors. It’s disruptive but ultimately recoverable. You could imagine this same scenario playing out in a law enforcement agency that contracts with Palantir, for example, as ties to the company increasingly become problematic in Democratic primary fights.

Now imagine something more extreme but entirely plausible. A president comes to power deeply opposed to the small number of firms that represent all the major players in frontier AI, perhaps because she made data center opposition a signature part of her campaign. Before the election, those firms announce — as is their right — that they’ll decline to work with that candidate if she wins. But this capital strike would leave an incoming government without options, particularly as citizens have come to expect AI-enabled service delivery. If the IRS is staffed to function with a significant chunk of its filing season call volume being diverted to an AI chatbot, for example, a new administration could find its first 100 days marred by the resulting chaos. The government could go after overt collusion in the above example, but what about the more ordinary and common case: a capital-intensive industry controlled by an extremely small group of people that simply has different political preferences or interests than the people in power?

This is not a technocratic problem. It’s a deep dilemma fundamental to liberal democracy, which has always struggled with the tension between individual liberty and the practical concerns of the state and society. Banning discrimination in public accommodations also impairs free enterprise and free association, but it is understood to be more desirable because it enhances the liberty of everyone else. Some monopolies (e.g., for utilities) are tolerated because astronomical infrastructure costs permit no alternative, but they are regulated more closely as a result. The original idea of the Defense Production Act was similarly reasonable – in a time of total war, it might really be necessary for the government to jump to the front of the line for steel that was being produced anyway. But generally these cases do not involve the compulsion of private action, in particular when that action is expressive.

In a real sense, finding the right balance between these competing interests is the hard work of democracy. Despite what unitary-executive fanatics would have you believe, getting it right requires broad political engagement from all of our institutions and not simply a pronouncement from the president. And, even if you did agree to compel action by a privately held company, would you really want to hinge national security on the work of anyone helping under duress?

The principal-agent problem we can’t solve the old way

The administration’s DPA gambit is a lousy, dangerous, and incoherent answer to this question, but it is not a lousy question. The more the government entwines the private sector in its day-to-day operations, the more difficult it is to ensure that the tail doesn’t wag the dog. This is the heart of it, and it’s easily obscured amid the storm over Hegseth’s tantrum.

We’ve slowly and painfully learned that to solve the principal-agent problem in government technology, we need to put a technically skilled person on the public’s side of the table. It took us a couple decades to reach this insight for regular software development, and we still struggle to act on it. Now, it may be on the verge of becoming obsolete as we get lapped by technology.

This is a bigger claim than “Anthropic should cooperate” or “the Pentagon was too aggressive.” It’s a claim that our entire framework for managing the government’s relationship with its technology vendors needs a radical overhaul, and that we need to get it right much faster this time.

No amount of DPA threats, supply chain designations, or contract renegotiations will fix a structural problem. Rather than threatening to compel industry to work for the government on the government’s terms, an administration genuinely concerned about this should be working with Congress to build something new.

One option would be a publicly-governed national lab for AI structured as a Federally Funded Research and Development Center, perhaps housed at a consortium of public universities across the country or patterned after the national laboratories. It could also look like a direct government capability along the lines of NASA’s early monopoly on space flight – an agency with a Senate-confirmed head, annual appropriations subject to congressional oversight, and all of the transparency requirements that follow from that. It could also look like a venture controlled by multiple branches of government simultaneously. The federal government was in the driver’s seat for many of the most important innovations of the modern age – the internet, GPS, space flight, etc. – and could be again.

This wouldn’t and doesn’t need to replace private AI vendors, but it would give the federal government an independent capacity to evaluate, test, and when necessary, replicate what the private sector is providing. Entities like the National Institute of Standards and Technology are already doing some of this work, but not necessarily with an eye towards replication and without the capacity to deliver service directly. In cases where it would be inappropriate or entirely too risky to deploy a private sector model, having a government-operated baseline would provide some measure of security for policymakers of both parties.

I want to be honest about what I don’t know here. A public AI entity would face enormous challenges: competing with the private sector for talent, keeping pace with a technology that’s evolving at breakneck speed, and operating inside a government procurement and personnel system that was not designed for this kind of work. It’s hard to imagine running an AI lab of any consequence when faced with constant government shutdowns, for instance, or one that has to navigate growing public backlash to data centers and the associated political incentives to tear things down. It would also potentially enable applications of AI on topics that no private lab would be willing to support but are nonetheless choices Congress and the President agree to direct.

But even if we never build a public AI lab, we need new governance models for the vendor relationship itself. If the fundamental problem is that the government can’t verify what’s happening inside the technology it depends on, then we need institutions designed to address that specific gap. This could mean mandatory interpretability standards. It could mean independent third-party auditing regimes with real teeth. It could mean extremely long-term contracts (say 10 years) entered into jointly by Congress and the president. It could mean moving deprecated, well-understood (but still useful!) models into the hands of the government when they “retire” as part of the normal business cycle. It could mean something we haven’t thought of yet.

Act before we’re stuck

We’ve seen a version of this movie before. The federal government’s transition to complex computer systems in the 1980s and 1990s happened without enough serious, careful thinking about vendor governance, and we ended up with exactly the captured, dysfunctional ecosystem we’re now trying to reform: legacy systems held together by contractors who can never be replaced, billions wasted on modernization projects that fail because the government doesn’t have the capacity to manage them, and the feeling that generations of policy professionals have that they aren’t actually in control of the government.

AI is going to turbo-charge this problem. The models are less auditable than traditional software. The market is more concentrated. The dependency, once established, will be harder to unwind. And the stakes, given what these systems will be asked to do, are considerably higher. If AI opens the aperture on disruption, it should also open our aperture for solution.

The DoD’s Defense Production Act debacle is the wrong answer. It’s inconsistent with the liberal values at the heart of the American tradition. It stifles private enterprise and innovation. It supposes that the government can arbitrarily compel action of private citizens without a single vote in Congress or take their property without due process. It does none of the careful balancing between the public’s basic interests in a government not reliant on the good will of a small handful of privately-held companies, the clear right of those companies to decline some work, and the practical challenges of day-to-day deployment. It is odious and wrong on so many levels that it’s not surprising the administration is having a hard time justifying their action.

But there is a glimmer of a good question inside of this whole affair: Who controls the technology that the government depends on, and how do we ensure adequate protection of the public’s interests when that technology is fundamentally opaque?

This has been one of the most important governance questions of the last decade and will only become more important in the next one. We need to start building the institutions to answer it now, before we wake up one day and find ourselves in a government so reliant on private firms as to barely be much of a “public” sector at all.

Gabe Menchaca is a Senior Policy Analyst at the Niskanen Center and, among many other things, is a former management staffer at the Office of Management and Budget and former management consultant. At Niskanen, he writes about civil service reform, the state capacity crisis, and other government management issues.

Share

Subscribe now

This essay originally appeared on If You Can Keep It, the Protect Democracy Substack.

In the timely Star Wars spinoff Andor, the young rebel Karis Nemik writes, “Tyranny requires constant effort. It breaks, it leaks. Authority is brittle.”

This is the lesson of history. Authoritarianism is not new. It has appeared throughout human civilization. But we live in a largely free world today because of Nemik’s observation: Authoritarianism has been — at least historically — unstable. It has relied on imperfect information, fallible human enforcers, and coercive systems that strain under their own weight. And because of that brittleness, people have repeatedly been able to correct away from tyranny and toward freer, more democratic societies.

But what if that were to change?

In his recent essay, “The Adolescence of Technology,” AI executive Dario Amodei opens with a different science-fiction reference: a scene from Carl Sagan’s Contact meant to illustrate a sobering idea — that societies capable of achieving immense technological power must also learn how to survive it.

Amodei, co-founder of the leading AI lab Anthropic, lists the threats we have grown accustomed to hearing about: human extinction, human subjugation to misaligned AI, mass unemployment. But he adds a danger to the list that needs far greater amplification and attention: the possibility that AI could undermine the historical dynamic Nemik describes, the very dynamic that has been democracy’s salvation. That AI could harden authority rather than expose it. That it could seal the cracks through which freedom has always eventually re-emerged.

In short, Amodei is warning that powerful artificial intelligence may be the first technology in human history capable of making authoritarianism permanent — not by violently overthrowing democracy but by eliminating the possibility of democratic counter-movements.

If that is correct (and, as Amodei acknowledges, there are still plenty of uncertainties that remain regarding both whether AI will achieve such advanced capabilities and the speed at which this would occur), we are not merely facing another technology policy challenge. We are approaching a civilizational threshold.

What’s more, while Amodei’s assessment largely looks to a future moment when AI has become powerful enough to be the equivalent of a “country of geniuses,” the harms caused by AI in the hands of authoritarians are all too real today. If we are to preserve the possibility of democratic pushback, we must act now.

Why authoritarianism has always failed

Authoritarian regimes have always appeared formidable from the outside. Yet their defining feature has been fragility.

They suffer from information problems. Fear distorts reporting; loyalty replaces truth. The “Dictator’s Trap” is that everyone around an authoritarian is afraid to deliver bad news and so provides him with misleading information that leads to overreach and error.

Authoritarians rely on human agents who hesitate, defect, or leak information and who have capacity constraints by dint of their very humanity. Such leaders must tolerate inefficiencies and often breed and encourage corruption to maintain control.

And because repression is costly — politically, economically, and psychologically — it tends to provoke resistance, fracture elites, and invite external pressure.

Even the most brutal regimes of the 20th century eventually broke. Fascist states collapsed under war. Communist regimes stagnated, splintered, or reformed. Military juntas gave way, sometimes suddenly, sometimes haltingly, to civilian rule.

The lesson is not that democracy is inevitable. It is that authoritarianism has never been able to fully close the loop. Something always leaked.

Amodei’s warning is that “powerful AI” could change this equation:

Current autocracies are limited in how repressive they can be by the need to have humans carry out their orders, and humans often have limits in how inhumane they are willing to be. But AI-enabled autocracies would not have such limits.

Authoritarian lock-in

Amodei identifies a convergence of sufficiently powerful AI-enabled capabilities that, taken together, threaten to eliminate the mechanisms through which authoritarian systems have historically failed:

• Total surveillance: Systems become capable of ingesting vast amounts of data on communication, movements, and behaviors so that there’s no escaping or hiding from the controlling apparatus of the state. Resistance can be identified and rooted out before it can bear fruit.

• Personalized propaganda: Persuasion is optimized at the level of the individual, leveraging intelligence on emotional states, social context, and psychological vulnerabilities that is collected by scraping data from digital (and potentially even real-world) interactions. Individuals can be manipulated into consenting to giving up their freedom, rights, and willingness to dissent.

• Autonomous weapons: The state has the ability to enforce suppression at scale through automated tools capable of responding with full force to every transgression and every minor act of dissent or organizing — and to do so potentially autonomously and thus unconstrained by the capacity and accountability limits of human bureaucracies.

• Strategic mastery: With a “country of geniuses,” a leader or entity bent on power could strategically outmaneuver any domestic or international opposition that is not similarly equipped, extending indefinitely the totality of their control.

Together, these capabilities point toward what can be described as authoritarian lock-in: a condition in which a regime does not merely suppress opposition but structurally prevents it from ever becoming effective enough to displace it.

The danger is that powerful AI could seal authority rather than strain it — closing the feedback loops through which societies have historically corrected themselves. Once these feedback loops are closed, the society never matures beyond its “technological adolescence”; it simply ossifies.

This concern differs in a significant way from concerns about human extinction or human subjugation at the hands of powerful AI. Serious as they are, those concerns are speculative and await the advancement of powerful AI, whereas the scenario of AI being used to empower authoritarian suppression has already arrived.

China is using AI to track Uyghurs with facial recognition, to identify potential dissidents with predictive policing, to suppress speech at scale with content moderation. The Chinese Communist Party is exporting this technology to other authoritarian regimes. And here in the U.S., the current administration is already implementing some of these tools.

The question that follows: What do we do about it?

Within the AI community and U.S. political community, there is already a dominant answer to this question.

Most leaders and thinkers closest to frontier AI — including Amodei himself — identify the primary danger as the possibility that China develops powerful AI first.

The reasoning is straightforward. China is already an authoritarian state. If it acquires powerful AI systems capable of total surveillance, predictive repression, and overwhelming strategic advantage, it would not merely entrench authoritarian rule at home. It could utilize its AI-enabled power advantage to extend its control abroad, potentially even achieving global hegemony under a totalitarian umbrella. Authoritarian lock-in, but without borders.

Fear of China’s global hegemony augmented by these tools often points to a single overriding imperative: The United States must beat China to the most powerful AI. The Trump White House has put it in those terms, declaring that “to remain the leading economic and military power, the United States must win the AI race.” On Capitol Hill, the message is even more explicit. Senate Commerce Committee Chair Ted Cruz has argued that “the way to beat China in the AI race is to outrace them in innovation,” urging policymakers to “remove restraints” that slow development.

Industry leaders echo the same logic: Microsoft President Brad Smith told lawmakers that the “number one factor” determining whether the U.S. or China “wins this race” is which technology is adopted globally, warning that “whoever gets there first will be difficult to supplant.” Former Google CEO Eric Schmidt has repeatedly warned that AI leadership will shape the global order. The result, as the Washington Post observes, is a “near-consensus” among senior officials and top executives that the U.S. must let companies “move even faster” to maintain its edge over China.

It is for this reason that Amodei himself describes advanced chip export controls as among the most important interventions available to democratic governments. The underlying theory is clear: If AI is dangerous, it is far more dangerous in the hands of an authoritarian state than a democratic one.

This assessment has profoundly shaped the orientation of the U.S. AI ecosystem.

The perverse dynamic it creates

The China-first framing is not wrong. But it is dangerously incomplete.

If beating China to powerful AI is the overriding objective, the logical conclusion is speed at all costs. Minimize regulation, remove friction, treat guardrails as liabilities, and view caution as complacency. This logic pushes U.S. firms to race ahead, consolidate power, and resist democratic constraints — all in the name of preventing authoritarian lock-in abroad.

But that same logic leaves us exposed to a second, equally serious risk: authoritarian lock-in at home. Amodei himself recognizes this risk in acknowledging that “a hard line” must be drawn “against AI abuses within democracies” with “limits to what we allow our governments to do with AI, so that they don’t seize power or repress their own people.”

Last year, one of us spoke with a founder of a leading AI lab. When we asked what worried them most, their answer was immediate: not China, but the possibility that extraordinarily powerful AI systems might come online while a leader with a deeply instrumental view of power controls the U.S. government.

This was not a partisan remark so much as a structural one. AI dramatically lowers the cost of surveillance, enforcement, and control. It centralizes authority. It weakens accountability. Those effects are dangerous in any system — but maximally so when wielded by leaders who already reject institutional constraints.

And as President Trump said just last month, “I don’t need guardrails. I don’t want guardrails. Guardrails would hurt us.”

The domestic authoritarian risk

The Trump administration largely accepts the China-first theory. It has embraced AI as a strategic asset and treated governance as an impediment to innovation. It has sought to remove guardrails in the name of competitiveness and national strength.

At the same time, our government is led by a president who has repeatedly demonstrated contempt for checks on his own power — attacking courts, undermining independent oversight, threatening political opponents, and praising strongman tactics. The administration has sought to expand surveillance authorities, suppress dissent, and consolidate executive power — directionally aligned with the very risks Amodei warns about. Ironically given a strategy ostensibly grounded in concern about what China would do with powerful AI, Trump has praised Xi Jinping’s leadership style, calling him a “brilliant” leader for “control[ling] 1.4 billion people with an iron fist.”

The uncomfortable reality is this: The problem is not only which nation gets powerful AI first but who controls that nation when it arrives.

If AI hardens authority, then leaders tempted by authoritarian methods, whether in Beijing or Washington, pose a similar structural threat. If we have one quibble with Amodei’s essay, it’s that it seems to underplay this domestic risk, perhaps because naming it explicitly carries costs for an American company right now. But that itself underscores the problem. And so we’re taking one of Amodei’s pieces of advice (“the first step is … to simply tell the truth”) and naming that risk more explicitly here.

Political economy makes this worse

Layered on top of this dilemma is a brutal political economy problem.

Frontier AI development is extraordinarily capital-intensive. Compute, data, and talent are concentrated in a handful of firms. Those firms command immense economic power — and increasingly, political influence.

As Amodei notes, AI is such a powerful economic and geopolitical prize that the risks of regulatory capture are intensified. Already, more than $100 million has been committed to an industry-backed super PAC committed to preventing regulation, and the pool of capital available to augment that is virtually bottomless. Democratic checks, already strained, risk being overwhelmed not just by urgency and fear of falling behind but by a huge campaign finance war chest.

The result is a convergence of concentrated private power and concentrated state power at precisely the moment democratic safeguards matter most.

Amodei acknowledges this problem but doesn’t offer a realistic proposal for solving it. If there’s a glaring failure in his essay, that is it.

The only durable solution

If AI can produce authoritarian lock-in, then preventing it requires more than choosing the “right” geopolitical winner or slowing development “slightly” (to use Amodei’s term) at the margins. It requires actively designing AI and its governance to advance democratic practices and values.

It will not be enough merely to anticipate how AI may harm democracy and impose penalties for its abuses and harms. Instead, the only durable solution is to account for AI’s risks to democracy in its development and deployment while using it to affirmatively make democracy work better.

That means AI systems that increase transparency rather than secrecy; that strengthen accountability rather than weaken it; that distribute power rather than concentrate it; that help citizens understand, deliberate, and participate rather than manipulate or surveil them.

These goals must be built into the design choices being made about AI now, at the market level, the corporate level, and the governance level. They need attention, investment, and innovation — today.

From insight to institution: the AI for Democracy Action Lab

This recognition is why Protect Democracy launched the AI for Democracy Action Lab.

What is missing in the AI ecosystem is not analysis, white papers, or technical benchmarks, nor efforts focused on existential risk to humanity (though those are important, too). What’s missing are institutions explicitly dedicated to defending democratic pushback — the capacity of societies to correct course when power is abused. And getting there means building AI in a way that both anticipates and wards off the risks of authoritarian lock-in while advancing and improving democracy and democratic commitments.

When one of us spoke to another top executive at a frontier lab recently and asked about the tenor of conversations in their c-suite and lunchroom about how AI might shape the battle between democracy and authoritarianism, their answer was, “What conversations?” While that might have been hyperbole to make a point, it illustrated an imbalance that Amodei’s essay nobly strives to correct. We must invest far more attention and energy on the risk AI poses to democracy and democratic reversal.

That’s what the Lab is designed to do by focusing on three fronts:

1. Defending democracy from AI-enabled authoritarianism through litigation, regulatory measures, and effective governance.

2. Ensuring AI contributes to a healthy civic media ecosystem.

3. Harnessing AI to strengthen democracy by creating products that augment democratic practices.

This work is not only for civil society and government. It is the real challenge for AI companies themselves. For Anthropic and others who take Amodei’s warning seriously, avoiding authoritarian lock-in cannot be a secondary concern. It must be a design principle.

The choice before us

Nemik was right: Tyranny has always required constant effort. It has broken because it is brittle.

As Amodei argues, powerful AI threatens to eliminate tyranny’s inherent vulnerability. It threatens to seal authority, automate control, and make correction impossible.

And this is not just a shoe we expect to drop at some point in the future. The AI of today is already redefining authoritarian capabilities inside and outside the U.S.

Whether AI will make democratic governance more capable than it has ever been or render it obsolete will not be decided by the technology itself. That question will be decided by us as AI’s shapers and wielders. We are the factor that determines whether action is taken before authoritarian lock-in becomes irreversible.

We have argued that the contest between democracy and authoritarianism is the test of our time, and it is now clear that one of the central fronts of that contest will be which side AI advantages.

It is up to us to get that right before it’s too late. The clock is already running.

Ian Bassin is co-founder and executive director of Protect Democracy.

Nicole Schneidmann is head of the technology and data governance team at Protect Democracy.

Share

Subscribe now

Image: Wikimedia Commons

Remember the “techlash”? A few short years ago, expectations were widespread that discontent with the Silicon Valley giants could result in major political and legal change. The Economist suggested that corporate break-ups, bans on mergers, utility regulation, and changes to content-liability rules could be in the offing.

Indeed, the European Commission and the Biden administration took some ambitious steps. The EU chastened Apple, Meta, Amazon, and Google with antitrust actions; more significantly, it passed laws setting conduct and content standards for large “platform” companies and imposed new data protections that created annoying pop-ups but also gave users more control over their information. The Biden administration also went to court, launching antitrust cases against Amazon, Google, and Meta, and a separate suit against Amazon that recently yielded a $2.5 billion settlement. Shifting from brakes to gas, Congress passed the CHIPS and Science Act to assert national control of a vital industry. Acting out of national-security motivations rather than social ones, Congress banned TikTok (until Donald Trump voided the law). Perhaps most consequentially, the social media giants face an avalanche of personal-injury lawsuits and litigation from state attorneys general in echoes of the 1990s cases that bludgeoned Big Tobacco.

Those are instances of government baring its teeth, changing how the tech giants handle data and tweaking the industry’s structure. But so far, it remains far short of a bite that would give the mass public a lasting grip on a sector that is transforming our culture, economy, and politics at breakneck speed. In the United States, the last two years have seen a dizzying rise of artificial intelligence free of any comprehensive policy framework to mitigate risks and encourage provision of public goods, with the Trump administration signaling it will pursue “minimally burdensome” regulation; the ever-deepening intrusion of social media into our lives, lawsuits notwithstanding; and a political realignment that has seen the Silicon Valley elite scrambling to get on the right side of a president with a penchant for crony capitalism.

To critics, the meager results reflect the reality that the techlash never ran as deep as its champions proposed. “It was never the people v. tech, it was a fight between media and tech,” writes the entrepreneur and investor Erik Torenberg. But half of Americans have consistently said tech should be more heavily regulated, and nearly 80 percent believe social media companies have too much power, according to a 2024 study from Pew.

Now, there are signs that a new wave of revolt is gathering. AI’s embodiment in grimly purpose-built data centers is drawing grassroots opposition from across the political spectrum. Deeper fears of AI doom are generating proposals for more serious regulation of the technology. The EU has already acted. And the ever-growing anxiety over the cultural acid leaking through our phones is leading many schools and parents to clamp down.

Will we get a political movement that is equal to the challenge this time?

Our new Hypertext forum asks why Big Tech is so hard to rein in, and how we might generate the cultural and political power to build a digital political economy that enhances our democracy rather than eroding it.

• Historian Jennifer Burns recalls that the last industrial revolution seemed no less overwhelming to those who lived through it and was only tamed after 60 years of effort — one that required cultural change as well as political organizing. “Someday,” she writes, “we might look back to Jonathan Haidt’s The Anxious Generation and place it in the same category as Upton Sinclair’s The Jungle, the expose that forever changed American factories.”

• Longtime journalist and editor Christopher Allbritton argues we are trapped in a “fragmentation flywheel.” Unregulated tech platforms exploit their users and deploy the profits to build successor platforms; those users then flee to the new place, restarting the cycle. Meanwhile, the tech barons’ power is not just economic and political, but narrative. As Allbritton notes, “The muckraking that forced Standard Oil to accept regulation doesn’t work when the monopolist owns the megaphone.”

• Niskanen Center Senior Fellow Adam Garfinkle argues that we have not yet fully recognized what we are confronting, not least because the technology is designed to obscure it. “The level and nature of addiction to digital devices is misunderstood and thus underestimated,” he writes, “and constitutes a species of Catch-22 paralysis new to the history of attempts to grapple with communications revolutions.”

• Finally, my colleague Brink Lindsey invokes the 19th century temperance movement in America to wonder if we could do something similar for media. “Just as the antebellum temperance movement changed attitudes about alcohol, using moral suasion to reveal it as inherently addictive and dangerous, so a modern-day movement needs to raise public awareness of the addictive spell that virtual experience can cast on us — and the cognitively compromised nature of that virtual experience.”

In his landmark book The Revolt of the Public, Martin Gurri explained that digital technologies are enormously powerful in generating outrage and protest, but much less suited to building durable movements. That insight applies to backlash against technology itself for all the reasons Allbritton and Garfinkle elaborate.

But as Burns points out, the problem of adjusting to an economic order that shifts social and material relations in ways that undermine community building and protest is not new. When factories began mushrooming across the landscape, it took much time and a good deal of bloodshed for laborers to fashion unions that could effectively advocate on their behalf.

In America, early industrial development spawned a class of “mechanics” who sought their champion in Andrew Jackson. But the Jacksonian dispensation proved better at conquering new land for settlement than at managing an industrial economy; Jackson’s successor, Martin van Buren, became known as “Van Ruin” for presiding over a deep depression. Later, as the mechanics gave way to a true industrial proletariat, the Knights of Labor combined a progressive solidarity across ethnic divisions with a radical politics and enjoyed spectacular growth, but eventually found itself supplanted by the American Federation of Labor, which notched wins with a more exclusive strategy focused on highly skilled workers. Only after decades did the Congress of Industrial Organizations develop the techniques to effectively represent workers in the full range of jobs within a single industry.

Countervailing movements are slow to develop during a time of rapid industrial changes because people need time to invent new systems of organizing. Along the way, they must develop not only a negative agenda — knowing what they are against — but also a positive vision that moves the clock forward, with the new technology, rather than backward, against it.

In his recent essay “Abundance for what?” Lindsey suggests that a far more ambitious version of the abundance agenda could provide the basis for such a positive vision. As Lindsey writes:

This wider cultural turn against the sordid excesses of online life could be the wave that carries that abundance idea from its current niche status — the preoccupation of technocratic elites — and transforms it into a genuinely popular social movement. The negative motivations are already in place. Fears of genuinely dystopian dangers have been awakened — and what’s more, the people most exposed to those dangers are our children, rousing our passions all the more.

The idea here is to appeal to ordinary people — not technology enthusiasts who thrill to the ingenuity and brilliance of the new and pathbreaking, but regular, risk-averse folks who tend to be suspicious of change because of their natural focus on holding onto what they’ve already got. To generate mass support for a resumption of large-scale progress in the physical world, you’ve got to hold out the prospect of big, tangible gains in ordinary people’s lives.

That something, Lindsey suggests, is not just a turn back to real-life community, but the prospect of making improvements back in the physical world that drive the cost of living down far enough to free people from the necessity of lifetime employment.

It’s a vision that may sound utopian in this era of low trust and low hope. But that underlying despair is precisely what we are trying to beat. We’ve done it before.

David Dagan is director of editorial and academic affairs at the Niskanen Center. Find him on LinkedIn.

Share

Americans used to drink like fish.

For an impressive example, look no further than the Father of Our Country. On September 14, 1787, near the close of the Constitutional Convention in Philadelphia, George Washington dined and drank at City Tavern as the guest of the “First Troop,” a cavalry corps that had crossed the Delaware with Washington and endured Valley Forge. And here, for the guest list of 55 gentlemen, is what they drank: 54 bottles of Madeira wine, 60 bottles of claret, 22 bottles of porter, 12 bottles of beer, 8 bottles of cider, and 7 large bowls of punch. No other details about the evening are available, but it seems safe to assume that a good time was had by all.

By the standards of the day, it doesn’t look like the evening was anything especially out of the ordinary. According to W. J. Rorabaugh’s The Alcoholic Republic, in the early 19th century Americans older than 14 averaged 7 gallons of 200-proof alcohol a year – well more than double the current U.S. average. In the 1820s, roughly half of all adult men were drinking at least six shots of liquor a day. Drinking water was widely considered unsafe so alcohol was served at every meal, liquor was cheap, and taverns served as centers of local social activity. Visitors from abroad regularly commented on Americans’ hard-drinking ways.

But by the middle of the 19th century, the situation was completely different. According to Rorabaugh, alcohol consumption had plummeted to under 2 gallons annually – below current usage rates. What could have happened to provoke such a dramatic change in the space of a couple of decades?

What happened was the American temperance movement, part of a much larger wave of moralizing and reform unleased by the religious revival of the Second Great Awakening. America had known religious enthusiasm since its earliest settlers, but by the early 19th century religiosity in the country was at a low ebb. The influence of the Enlightenment among the well-educated could be seen in the large number of deists among the Founding Fathers; meanwhile, as the country expanded westward, its frontier was wild and largely unchurched. In an 1822 letter to a friend, Thomas Jefferson predicted, “I trust there is not a young man now living in the Unites States who will not die a Unitarian.”

But even as Jefferson was writing those words, a spiritual revolution was in the offing. As I described it in The Age of Abundance:

During the early decades of the nineteenth century, a surge in evangelical fervor that came to be known as the Second Great Awakening swept over the young nation. In New England, Nathaniel Taylor and Lyman Beecher (Harriet’s father) spearheaded the development of “New School” Presbyterian theology. Charles Grandison Finney honed soul winning into a science with his massive revival campaigns in New York and the Midwest; Finney’s ministries were credited with achieving hundreds of thousands of conversions. Camp meetings in the South and the Appalachians sparked explosive growth of the Baptist and Methodist denominations. Between 1780 and 1820, Americans built 10,000 new churches; over the next four decades, they added 40,000 more.

The revitalized American Protestantism that emerged from the Second Great Awakening broke decisively with the Calvinist past. Specifically, it rejected Calvinism’s central dogma of predestination and asserted the individual’s free moral agency… In its more optimistic strains, the new American religion went beyond mere free will to claim the possibility of human perfectibility… “If the church will do all her duty,” Finney proclaimed in 1835, “the millennium may come in this country in three years.”

This massive mobilization of spiritual energy, fired by a vision that exalted bourgeois respectability as Christian virtue, inspired a series of moral crusades. Most noteworthy was the rise of the abolitionist movement, as anti-slavery sentiment shifted from support for gradual resettlement in Africa to campaigning for the immediate and unconditional freeing of slaves. William Lloyd Garrison started publishing The Liberator in 1831, and helped to found the American Anti-Slavery Society two years later. Within a mere five years, the Society had over a thousand local chapters and 250,000 members.

Other campaigns for moral uplift proliferated – against gambling, bear-baiting, and cockfighting; against public whipping and mutilation for crimes; against premarital sex (in New England, it’s estimated that one bride in three was pregnant in the late 1700s; by 1840 it was one bride in five or six); and last but not least, against the scourge of alcohol. The American Temperance Society was founded in 1826, and within a dozen years it boasted some 8,000 local groups and over a million members.

The key innovation of the new temperance movement was to see alcohol as inherently dangerous. Although drunkenness was always considered sinful, as was gluttony, alcohol was seen as no more blameworthy than food; the fault lay entirely with those who overindulged. “Drink itself is a good creature of God, and to be received with thankfulness,” declared the Puritan clergyman Increase Mather back in 1673, “but the abuse of drink is from Satan.”

Lyman Beecher, one of the co-founders of the American Temperance Society, delivered a series of sermons on “intemperance” in 1826, which were then published and went on to enjoy brisk sales for decades. He disputed the common opinion that someone who drinks regularly with no signs of intoxication is doing nothing wrong: “Whoever, to sustain the body, or invigorate the mind, or cheer the heart, applies habitually the stimulus of ardent spirits, does violence to the laws of his nature, puts the whole system into disorder, and is intemperate long before the intellect falters, or a muscle is unstrung.”

“The effect of ardent spirits on the brain, and the members of the body, is among the last effects of intemperance, and the least destructive part of the sin,” Beecher maintained. “It is the moral ruin which it works in the soul, that gives it the denomination of giant-wickedness.”

At the heart of alcohol’s evil, according to Beecher, is its effect on the will – that is, its addictiveness. An ardent abolitionist (recall he was the father of the author of Uncle Tom’s Cabin), he argued that submission to alcohol was another form of slavery:

The demand for artificial stimulus to supply the deficiencies of healthful aliment, is like the rage of thirst, and the ravenous demand of famine. It is famine: for the artificial excitement has become as essential now to strength and cheerfulness, as simple nutrition once was. But nature, taught by habit to require what once she did not need, demands gratification now with a decision inexorable as death, and to most men as irresistible.

We execrate the cruelties of the slave trade—the husband torn from the bosom of his wife—the son from his father—brothers and sisters separated forever—whole families in a moment ruined! But are there no similar enormities to be witnessed in the United States?

Every year thousands of families are robbed of fathers, brothers, husbands, friends. Every year widows and orphans are multiplied, and grey hairs are brought with sorrow to the grave—no disease makes such inroads upon families, blasts so many hopes, destroys so many lives, and causes so many mourners to go about the streets, because man goeth to his long home.

The Jacksonian-era temperance movement radicalized over time, calling for total abstinence rather than mere moderation and seeking legal prohibition to ban alcohol production and sales. But its greatest achievements came from its incrementalist moral suasion. Beecher started with his own profession, speaking out against drinking by the clergy. He moved on to campaign against employers serving liquor on the job: “It is not too much to be hoped, that the entire business of the nation by land and sea, shall yet move on without the aid of ardent spirits, and by the impulse alone of temperate freemen.” We can see in such tactics an analog to the “clear and hold” approach to anti-insurgency: pacifying the threat in some particular location, creating conditions that ensure the threat will not return, and using the momentum from such partial victories to carry the fight to other domains.

(I’ll note that the temperance movement petered out with the outbreak of the Civil War, but was reconstituted in the 1870s. This later temperance movement culminated in Prohibition – oops, a bridge too far. So I’m focusing here on the more modest, and more effective, earlier movement.)

The Second Great Awakening more generally, and the antebellum temperance movement in particular, hold important lessons for us today. First, they show us that – contrary to a pessimistic strain of opinion common on the right – a society’s “moral capital” is not just some inheritance from the pre-modern past that is inevitably drawn down as the old traditions fade. Broad-based moral regeneration can occur under the conditions of modernity, and it can be rapid and dramatic. More specifically, the temperance movement shows us how a free society can respond to the challenges of addictive activities that subvert individual autonomy. In a free society, we generally presume that people should be allowed to do what they want. But informed by the distinction between liberty and license, we recognize that sometimes we face a conflict between our “first order” and “second order” wants: we may simultaneously desperately want a drink and desperately want to be free of that desire. The temperance movement shows that education combined with moral suasion – raising awareness of the threat posed by some addictive activity, and liberally wielding praise and blame to incentivize right conduct – can be effective in keeping self-defeating abuses of freedom in check.

Today we are faced with a number of deepening social ills as a result of another species of intemperance – overindulgence in the consumption of mass media. According to Nielsen data, American adults averaged a little more than 11 hours a day consuming media in the first quarter of 2018: 4 hours, 46 minutes watching TV; 3 hours, 9 minutes using a smartphone or tablet; 1 hours, 46 minutes listening to radio; 39 minutes using the internet on a computer; and 39 minutes playing games. As for young people, the New York Times reports that, as of 2019 (i.e., prior to the Covid-related school closings that drove numbers even higher), the average amount of time staring at screens stood at 4 hours, 44 minutes a day for tweens (ages 8 to 12) and 7 hours and 22 minutes for teens (ages 13 to 18). It’s no wonder that the word “binge” now more commonly refers to TV viewing than to going on a bender.

These aggregate figures are eye-popping, and on their own reveal that all is far from well. Even if all the content consumed were wholesome and edifying, the sheer bulk of the time expended suggests serious problems of opportunity cost. And of course, we know that a great deal of media content is mental junk food at best and at worst can badly distort our sense of reality – whether by inundating us with images of airbrushed physical perfection that accentuate our own inadequacy, or by spreading conspiracy theories and other arrant nonsense. Accordingly, it is now widely accepted that current media habits are unhealthy and that screen time is associated with all kinds of negative side-effects – including obesity, loneliness, anxiety, and depression.

Let me make clear that I am far from puritanical about such matters. Over the course of my lifetime I’ve watched great gobs of TV, much of it eminently forgettable dreck, and these days I spend the bulk of my waking hours staring at a screen – mostly for work, but with lots of time-killing frolics and detours along the way. And yet it hasn’t stopped me from reading widely and deeply, traveling all over the world, and enjoying deep and abiding personal relationships. Indeed, shared media experiences with friends and loved ones have served to strengthen our bonds and now make for some of our fondest memories.

So I want to home in on four particular aspects of contemporary media consumption that I believe are especially problematic and in need of a concerted remedial response: (1) social media; (2) solitary consumption; (3) crowding out of deep literacy; and (4) media coverage of politics.

I won’t say much about social media, as so much has already been written and the issues are already so familiar. Social media sites promise to bring us together and strengthen our connections to friends and family, and they do sometimes deliver on that promise. It’s typical for those of us who were adults when Facebook hit the scene to have a longish list of folks from our past with whom we’ve reconnected; and for parents with kids, keeping grandparents, aunts and uncles, and family friends in the loop for all their various milestones is now almost effortless. Yet it has become painfully clear that social media has a deeply troubling dark side, especially for younger people. As I’ve written about previously, the past decade or so has seen a spike in mental health problems among adolescents, and the fact that this has occurred just as social media usage was taking off does not appear to be a coincidence. At the heart of the problem seems to be social media’s encouragement of “upward social comparison” – judging ourselves on the basis of a flood of highly selective and manipulated images and text designed to show people in the most flattering possible light.

This and other problems caused by social media aren’t just unfortunate side effects: the leading social media sites are designed to be unhealthy. They market themselves as helping people to connect, but their overriding priority is to keep us connected to the site no matter what. “Like” buttons, in particular, drive engagement through the dopamine hits we get when people respond positively to our posts. But the existence of those buttons turns social media into a ruthlessly precise, nonstop, global popularity contest, in which the vast bulk of us are destined to come out on the short end of the stick.

In much of life, what we’re doing matters less than whom we’re doing it with, and that’s true of consuming media as well. TV programming can be as shallow and mindless as you like, but watching and laughing and commenting with other people can make for an enjoyable and enriching social experience. Video games may be a huge time suck, but whiling away those hours with buddies isn’t time wasted in my book. The problem is that TV and the internet lend themselves so easily to solitary enjoyment – since they provide a kind of ersatz companionship. Think of old people living alone with the television constantly on in the background to chase away the silence; think of teenage girls sitting in their bedrooms, scrolling through Instagram for hours. Media usage makes it easier for people to withdraw from the world and avoid the real human contact they so badly need.

My Niskanen colleague Matt Yglesias wrote a good Substack post on this point titled “Sitting at home alone has become a lot less boring, and that might be bad.” Here’s a sample:

I’m obviously not going to give up the convenience of streaming home video, and neither are you.

But it is true that if my home viewing options were worse, I’d probably be inclined to ping friends a bit more frequently to see if they want to go see a movie. And those friends would probably be a bit more inclined to say yes to such invitations. They’d also probably be a bit more inclined to ping me about going to the movies…

And I’m inclined to say we’d probably all be better off for it.

To sit home, alone, and stream (which, to be clear, I do a lot!) is fun and easy and convenient. But it strikes me as potentially fun and easy and convenient in the same sense that it’s easy and convenient to not exercise or fun and easy to gorge yourself on Pringles. We’re weak creatures and can be easily tempted into patterns of behavior that we would reject if we had the ability to program our short-term behavior to align with our long-term goals.

It’s not just logistical convenience that links media consumption to self-isolation – it’s psychological, emotional convenience as well. Real, flesh-and-blood people are sometimes moody, sometimes demanding, sometimes boring, sometimes annoying; real people can’t interact with each other over the long term without some friction. By contrast, TV characters never argue with you, porn stars never reject you, you can restart games at the precise spot you messed up, and you can mute yourself and disappear from social media the second anything gets boring or uncomfortable. The more habituated you get to this undemanding, frictionless substitute for genuine social interaction, the more difficult and burdensome the real thing can start to seem. (For more on this, read my earlier essay “Choosing the Experience Machine,” and in particular check out the excellent Freddie deBoer essay I discussed there titled “You are You. We Live Here. This Is Now.”

Excessive media consumption isn’t just making us fat, lonely, anxious, and depressed; it’s making us stupider as well. I’ve also written about this before, so I won’t belabor the point; instead let me once again point you to my friend Adam Garfinkle’s essay “The Erosion of Deep Literacy,” as well as Neil Postman’s decades-old but still well-aimed jeremiad Amusing Ourselves to Death. The bottom line is that the cognitive culture of print is much intellectually deeper, richer, more complex, and more demanding than the cognitive cultures of either television or the internet. Reading books requires sustained focus and attention, logical thinking to follow and assess elaborate exposition and argument. TV viewing, by contrast, is passive, unfocused, carried along by moving images and emotional cues, while internet scrolling confronts our fractured, overloaded attention with jumbled mishmashes of disconnected information and hot-button stimulus. I believe that we can see our passage to a post-literate culture in the 21st century phenomenon of the “reverse Flynn effect”: after a sustained rise in raw IQ scores in advanced countries from the 1930s through the 90s (discovered by IQ researcher James Flynn and named after him), more recently the trend is going in the opposite direction. In my view, the original Flynn effect reflected our adaptation to the more complex and cognitively challenging environment of technology- and organization-intensive industrialism; brains trained on Tik Tok videos and emojis are generally not going to be as capable at maintaining focus and reasoning abstractly.

Finally, our current media environment is flatly inconsistent with the healthy operation of democratic politics. Here again, this is ground I’ve covered already. The fundamental problem is an ineradicable conflict between the profit motive and democracy’s need for a well-informed public. This conflict can be managed well enough when media competition is restrained (as it happened to be, in both newspapers and broadcast, during the first three-quarters of the 20th century); but when, as now, there is intense competition for readers’ and viewers’ attention, media providers will inevitably stoop to conquer. That is, they will be driven to transform the coverage of public affairs into another species of entertainment, offering up hot-button sensationalism, cartoonish contrasts of black and white, and an overriding focus on the competitive drama of who’s up and who’s down instead of on the substance of governance.

More specifically, it remains underappreciated the degree to which the rise of contemporary authoritarian populism here in the United States was the creation of a small group of media entrepreneurs. First talk radio, then Fox News, then the right-wing internet used technological innovation to reach audiences not aligned with the centrist liberalism of the old media establishment; they courted and built that audience by pandering to its prejudices and offering up a steady red-meat diet of outrages and demonization. Decades of such pandering created the contemporary GOP base, totally detached from consensus reality and demanding ever-more lurid political theater. Once Trump came on the scene, mainstream media took the bait and, cashing in on the intense media interest that right-wing demagoguery generated, cast themselves as leaders of the “resistance” and offered counter-theater to thrill their progressive constituents. CBS Chairman Les Moonves said the quiet part out loud when he admitted of Trump’s candidacy, “It may not be good for America, but it’s damn good for CBS.”

Here then are my proposed targets for a 21st century media temperance movement: (1) social media use, especially among young people; (2) solitary media consumption; (3) the decline in reading the printed page; and (4) political infotainment. While these are the specific abuses that the movement would seek to remedy, the broader campaign should be to instill in the public, or at least a critical mass of the educated public, a healthy suspicion of virtual experience in general. Just as the antebellum temperance movement changed attitudes about alcohol, using moral suasion to reveal it as inherently addictive and dangerous, so a modern-day movement needs to raise public awareness of the addictive spell that virtual experience can cast on us – and the cognitively compromised nature of that virtual experience. Every time we switch off the real world to tune in the mediated one, we should feel a little twinge of discomfort, as if we’d stepped into a disreputable dive bar in the middle of the day. We should be on our guard.

With regard to the specific abuses I’ve enumerated, the movement should aim to be ambitious, starting with changing attitudes and moving on to changing actual habits. Age limits for access to social media should be raised and enforced; parents should see teen phone use in the same light as teen drinking. Initiatives should be launched to design non-addictive social media sites that respect user privacy and autonomy. The general sense of guardedness toward the virtual world should be heightened when entering it alone; excessive solitary media consumption should be regarded as worrisome, while family viewing time and neighborhood watch parties should be celebrated and encouraged. Book reading should be promoted for all ages, with sponsorship of book clubs and public book-reading events and public service messages featuring media celebrities as class traitors. Online “political hobbyism” should be stigmatized as vulgar and creepy; cable news viewing of whatever ideological stripe should be actively discouraged through organized boycotts.

Could such a thing actually come to pass? The raw materials for a media temperance movement are already lying around at hand: there is widespread awareness of the general problem, and there are already numerous groups actively working on many of the movement goals I mentioned above. Whether these disconnected responses can mobilize into a coherent and energetic movement is anybody’s guess. But this much seems clear: it’s hard to picture a dramatically better society without a dramatically improved media environment.

This essay first appeared on Brink Lindsey’s Substack The Permanent Problem.

Even well-intentioned and well-organized communities have trouble fixing problems for which they lack a general consensual idea of cause. When it comes to social and political complaisance in the face of a tech-driven rewiring of the social order, our problem is not a paucity of explanations — quite the opposite. Just a short list of reasons commonly cited for the near-unchallenged ascendancy of the tech firms would include:

• the cumulatively deep depletion of social trust—Robert Putnam wrote the seminal essay for Bowling Alone all the way back in 1995, thirty years ago!—which makes self-propelled community organizing more difficult to start and less robust to sustain even when it can be started;

• the shocking erosion of deep literacy, which depletes social reservoirs of empathy and keeps knowledge of social media harms limited to mostly college-educated adults, whose cognitive styles are frequently off-putting to others in a burgeoning populism-friendly cultural environment;

• the generally numbing and reality-blurring effect of a culture in the throes of an onrushing entertainment-technology singularity, what David Foster Wallace called Total Noise in his 1996 novel Infinite Jest, defined by a situation in which “the marginal cost of an additive unit of entertainment trends to zero”;

• the Gattling-gun level of spectacle-besotted distractions coming from the White House alone;

• and more besides.

Most of these factors, however, are peripheral or additive to the main reason for our passivity, which breaks down into two parts: Socially, our cyberaddiction is more widespread and more pernicious — and thus more demobilizing — than is generally appreciated. And politically, the superhuman returns to scale generated by information technology have created an industry that rivals government itself in revenue and power. Neither part of the answer, alas, is simple or obvious to most observers, and that works to compound the problem.

Social media is only one facet of a larger and more generic technological juggernaut — the cybernetic revolution, or what I call the cyberlution. Norbert Wiener coined the term “cybernetics” in 1948 to mark the burgeoning study of information flows in complex systems. Cybernetics is now in digital mode and on the portal of a vast artificial intelligence expansion. We are hurtling toward that future in a culture that has only the faintest understanding of how the cyberlution has already transformed our politics — and our minds.

The doom loop of digital distraction

The key to the passivity most observers espy on the socio-cultural side is that the level and nature of addiction to digital devices, especially smartphones but also gaming systems and certain other gadgets, is misunderstood and thus underestimated. It constitutes a species of Catch-22 paralysis new to the history of attempts to grapple, legally and otherwise, with earlier communications revolutions, including those pertaining to telephony, photography, radio, and television.1 It is important to understand why this is before proceeding further.

What we observe as inherent to contemporary digital communications technology is partly old and partly new. What is old is two-fold. As Charles Horton Cooley, arguably the father of American sociology, argued in a seminal 1897 essay entitled “The Process of Social Change,” communications techniques feed what he called the “social mind,” that cloud-like normative structure in every culture that individuals are socialized into, influence only at the margins usually, and then ultimately pass out of. In that essay Cooley coined the term “social media,” which only hit the big time, beyond the reach of sociology professors, in somewhat altered meaning about a century or so later. And as Harold Innis put it at the portal of the television era in 1951, it is very nearly a law of social dynamics that all “sudden extensions of communication are reflected in cultural disturbances.”2

So much for what is old, not that most people actually understand any of this despite its hoary pedigree. What is new about the digital phase of the internet era can be expressed with reference to just three words: hyperconnectivity, disintermediation, and anonymity.

• Hyperconnectivity: The size and transmission speeds of today’s digital communications networks are orders of magnitude larger and faster than before, and their interactive transmission capacities at scale are unique.

• Disintermediation: Also unique, essentially no filters exist in the United States on digital communications transmissions, whether professional or governmental, except the very meek and limited ones designed for liability protection applied by the tech corporations themselves.

• Anonymity: It follows from radical disintermediation that senders of communications need not, often are not, and usually cannot be accurately identified by recipients of communications if senders so desire. This authorial ambiguity is characteristic of the whole digital internet era well beyond social media, and will only deepen with the rise of artificial intelligence — and the deepfakery we have witnessed thus far, while hardly trivial, will look like child’s play in a few years if current trajectories are left to unfurl unhindered.

The cyberlution has thus teleported us into an information environment of hyper-connectivity with rapid feedback loops, disintermediation that shatters quality control, and anonymity that shreds accountability for content. These features have created an online world of attention-seeking in the form of spectacle — a world, in short, that has gotten most of us addicted to distraction.

This cyberlutional addiction, quite aside from affecting many more people than earlier communications technologies, is different and more pernicious than most realize.

Most addictions are relatively easy for the sufferer to detect, even if he or she chooses to deny them; but once detected and confronted, they can often be treated and overcome. A gambling addict, for example, can see evidence of addiction in a dwindling bank account and an accrual of debt, and often enough in an array of riled up family relationships. Someone addicted to distraction, on the other hand, has a harder task. Timelines disappear for lack of any solid empirical referent that something is amiss, and as soon as a rare sign of a problem begins to dawn on the victim, whoosh, it’s gone, thanks to the next distraction. It short, addiction to distraction describes a virtual closed loop with no easy way out. It certainly does not help that so many people are now affected that addiction’s symptoms seem normal, even though they are anything but. We have long defined addictions as anomalous behaviors, so we are baffled into silence when they turn out to be as common as they have become.

This is important. The origin of any and every addiction is an illusion, a kind of fiction that some marvelous pleasurable reward may abide in one’s future. A person sets off in pursuit of this reward only to find that the faster he or she runs the more elusive the reward becomes. With substance addictions the usual response is to use more of the substance to sustain the same level of pleasurable hope that the ultimate reward can be grasped. With behavioral addictions like gambling, thrill seeking, and sex the usual response is to escalate the behavior in question. Since the ultimate reward is fictional, however, it will never be attained no matter the level or pace of the pursuit.

The same goes with addiction to digital technology, but differently, because the fiction is honed to near statistical perfection by algorithmic teasing. As Jaron Lanier put it, “The algorithm is trying to capture the perfect parameters for manipulating a brain, while the brain … is changing in response to the algorithm’s experiments.” But, continues Lanier, “because the stimuli from the algorithm doesn’t mean anything, because they are genuinely random, the brain isn’t responding to anything real, but to a fiction. That process — of becoming hooked in an elusive mirage — is addiction.”3

So for any practical purpose, all the elements of cyberaddictions are packed into an individual’s headspace, making them more insidious than conventional addictions even as they tend to be both less dramatic, less unhealthy in a manifest bodily sense, and — again — much less socially anomalous.

Moreover, cyberaddictions do not settle into a grind in the same way as substance and ordinary behavioral addictions. They are ever novel and endlessly attractive. One reason is that the push notifications that sparkle and the dings that resound from iPhones do not anticipate only one class of rewards. They can signal a new email, or a new text message, or a new YouTube video, or a calendar notification, or a recorded voicemail — and many if not most of these possibilities raise the expectation of a social connection, a force that in other cases arguably helps to navigate out of addiction. Even incidental, passing social connections seem to be more salient psychologically than most of us realize. So unlike the treadmill psychology of substance addictions in particular — that anticipatory, joyless “oh no, here I go again” feeling — the thrill is not gone (apologies to B.B. King), is never gone, from cyberaddictions.

This helps to explain why so many are in denial about the nature and dangers of cyberaddictions. They do not disable us as readily or as rapidly as do substance and behavioral addictions. They are subtler. They even let us remember the benefits of the technology, which can be real if the user manages to remain the master and not become the slave, remain Pavlov and not become the dog. But confidence that we will always be the master, and hence the too easy assumption that most others probably will be too, is a fool’s wager if ever there was one.

Look around, and keep in mind that addictions permanently change the brain, literally reshaping the dopamine dispersal network. When one reads descriptions of thorough-gone addicts who have plunged into the quasi-metaversal world of unbundled, short-form social media, it is hard to imagine soft landings that could reintroduce such mentality-disabled people to reality as it actually is.

More important to society at large, even in the majority, we generously assume, of non-extreme cases, social media immersion and the general lure of being online has shot the attention spans and quality attention capacities of tens of millions of Americans to hell. It has radically reduced the readiness of most students (and not only students) to read books and write their own essays— in other words, to learn how to think, manage time, and constructively plan on their own, without crutching on a ChatGPT bot.4

It has worsened our society’s alienation from nature and made us more, not less, lonely and isolated from each other, thus magnifying already sharply eroded social trust.

It has, thanks partly to its anonymous nature, helped to mainstream regressive zero-sum attitudes and expressions of hatred that have contributed to political polarization and incivility. It has also stimulated extreme and ahistorical debates, deranging our stock of knowledge and experience about scores of vital issues.

It has abetted the sexualization of children and the growth of conspiracy theories and cults, simultaneously pushing kids toward adulthood and adults toward childishness.

It has magnified the cognitive gluttony that is everywhere in the culture, further enshrining entertainment as the ultimate goal of an otherwise aimless civilization.

It has furthered the decay of moral reasoning and self-discipline, and accelerated the nadir of the Abrahamic faith communities that imbued American society with both, for there is little those communities can do to compete successfully with the graphic fantasy spectacle of the attention economy.

Over a mere few years, it laid the foundation for a fantasy-inflected, reality-television presidency, duly and fairly elected now twice. As a result, it has put the future of the United States as a constitutional, classical-liberal democracy at risk and, with it, arguably jeopardized the security of the entire global commons.

And in combination with other factors it has given rise to an immensely powerful digital oligopoly — an oligopoly of a very large scope and largely novel nature that even Robert Michels’ “iron law” never imagined and Theodore Roosevelt’s “malefactors of great wealth” barely touches. This oligarchy has arisen largely by dint of the Net Effect — and so we are brought willy-nilly to the second part of our answer.

The Net Effect, cost disease, and the new corporatism

While the critical attention social media has attracted is certainly justified, it tends to crowd out perception of other harms flowing from the broader technological stream of the digital revolution. Clearly, the ambit of the internet’s uses and hence its cultural influence since the world wide web went live in 1991 is far wider than its social media dimension.

This distinction is important because it turns out that the failure of U.S. politics to land any punches as it is beaten senseless by social media turns on these other, less widely recognized facets of the cybernetic revolution. We now behold a new and insidious form of corporatism in the Trump 2.0 administration, the direct antithesis of the ur-definition of classical liberalism, which separated economic power from political/coercive power. Today, the would-be looters of the domestic (and international) commons are inside government itself, and are largely directing its efforts, as opposed to ostensibly being regulated by it. Just as the Trump administration has switched sides, shockingly to many, in global geopolitics, it has also switched sides in terms of classical political economy definitions and functions. That is, or should be, even more shocking still. Most obviously, Elon Musk and his acolytes were put in charge of the henhouse, and though their initial rampage has ended, their influence lingers throughout a federal government that spends billions on contracts with the mogul; Trump illegally voided the TikTok ban to arrange a marriage that benefited his donor Larry Ellison and other supporters; and an assortment of crypto bros has provided the president with staggering wealth in exchange for his regulatory benedictions. In other cases, tech and tech-adjacent firms appear to be the willing victims of extortion, as in the case of the government’s claims on shares or revenues from Nvidia, AMD, Intel, and a trio of rare-earth miners.

How, in brief, has this happened?

The cyberlution has sired the Net Effect, a structural shift in American economic life that has literally outsized implications for our existing civil society and political institutions.

The Net Effect refers to an inherent characteristic of information technology innovation, especially in digital form: It aggregates a wide range of human transactions that were formerly more dispersed. Technology scales, and cybertech scales better due to its very nature, than any previous information science-cum-communications technology ever has. But human community and empathy do not similarly scale, nor do scientific-rational cognitive templates, at least not in the current American entertainment-centric culture. The result of this core characteristic of the Net Effect has been the relatively rapid transformation of American economic life, reshaping, and fragilizing, a generally affluent political economy in ways few understand.

As general observations go this one is at base not new. The relationship between advancing technology and advantages to scale in most things — manufacturing, banking, trade, investment, and more — is a staple insight into the core dynamic of the industrial revolution. It resounds in David Landes’ 1969 masterpiece The Unbound Prometheus, for example, but perhaps its most insightful succinct description is one from 1955, by John von Neumann:

In all its stages the industrial revolution consisted of making available more and cheaper energy, more and earlier control of human actions and reactions, and more and faster communications. Each development increased the effectiveness of the other two. All three factors increased the speed of performing large-scale operations — industrial, mercantile, political, and migratory. … Since most time scales are fixed by human reaction times, habits, and other physiological factors, the effect of the increased speed of technological processes was to enlarge the size of the units — political, organizational, economic, and cultural — afforded by technological operations. That is, instead of performing the same operations as before in less time, now larger-scale operations were performed in the same time.

Because, as many have noted, the cyberlution substitutes machine power not for human muscle but for selected human brain functions, the result is an exponential elaboration of von Neumann’s insight. If we substantially substitute machine-brain functions for human ones, especially in communications, we lift many of the limitations of “human action times, habits, and other physiological factors” from organizational processes, manufacturing and management processes alike. Indeed, as von Neumann later added, “. . . improvements in control are really improvements in communicating information within an organization or mechanism. The sum total of progress in this sphere is explosive.” Explosive, indeed: von Neumann saw at least dimly in 1955 much of what 2025 would look like.

The explosive progress von Neumann foresaw takes us far beyond descriptions of technology’s relation to scale during and after the industrial revolution. It multiplies advantages to scale dramatically, incentivizing gigantism decisively as an organizational goal. It thus transforms our intellectual inheritance of capitalism, since supply and demand no longer match up invisible-hand-like to produce price-points from massively dispersed economic agents engaged directly, without technological mediation, to create markets. Rather, digital information technology — which, again, is prior to and has little directly to do with social media — inherently conduces to oligopolistic, potentially even monopolistic, conditions.

In fact, the enormous size and swath of the tech companies are functions of the technology itself. Precisely because there is neither a product nor much of a human-devised service, output can increase instantly with the algorithmically-induced demand. Until recently, the capital investment required to provide the underlying computing power was a trivial problem; adding a server is much easier than adding a factory. Only with the extraordinary computing demands of AI has capital investment begun to represent a potential limit.

The Net Effect ramifies across the economy as nondigital businesses take advantage of the new technology to hoover up customer data, find new clients more quickly, and pursue ever-more complex financial wizardry. The resulting efficiencies could in theory be passed on to the consumer, but in key industries that are either oligopolistic, otherwise supply-constrained, or heavily regulated — including healthcare and housing — large providers instead use them to extract more profit. Even major grocery stores in cahoots with Instacart and various data-management firms — all far smaller than the truly major digitech corporations — have been using algorithms to sort buyers into categories that result in different people paying different prices for the same goods at the same time at the same store. This amounts to an ingenious (and not yet illegal) means of effecting wealth transfer from individual consumers to huge corporations’ managers, investors, and shareholders. And rather than just providing a tool for already-big service providers, the ease of monetizing data and the asymmetric advantages of complexity actively incentivize consolidation. The Net Effect thus aggravates what is already a nasty and novel case of cost disease.

So now finally, why has the U.S. federal government not reacted more assiduously to the harms posed by social media? Quite aside from the absence of pressure to do from below on account of the generally immobilizing nature of cyberaddictions, the fact is that the major digitech corporations, in informal cahoots in this regard despite their competitive relations in other regards, have more influence over the federal government than the federal government has over them. The combined annual revenue of the Big Five alone for 2024 summed to about $1.68 trillion — about double the size of the U.S. defense budget for FY 2024. The Big Five plus a few other huge tech corporations like Oracle account for, at last count, about 34 percent of U.S. stock market capitalization. Even the Treasury Department can’t readily touch that.

This situation puts the government on the losing end of a somewhat unusual, and hence opaque to many, collective action problem. All the Big Five, plus Oracle and Palantir and so on down the scale line, oppose regulation of their industry, all strenuously oppose any antitrust actions aimed their way, and they all share the same reasons for both. They are all also massively deep into government contracts and congressional campaign contributions. Compared to the federal government as a whole, they know what they want on what matters most to them, but the government—which we imagine as unitary but which in fact is a highly fragmented collection of public organizations—is all over the place, unable to concert any kind of coherent policy toward Big Tech.

This asymmetrical arrangement biasing policy in a generally passive direction held for years before January 20, 2025. Now, as noted, the MAGA Republican administration is actively abetting what amounts to a new corporatism in which digital mega-corporations are first-tier players. This accounts for why new techno-anarchist corporatists in the MAGA camp have in some cases mounted and in others supported efforts to intimidate those monitoring the internet for misinformation.5 Meanwhile, MAGA spinmeisters try to persuade their largely populist, post-literate constituency that the accursed monstrous bureaucratic/administrative state run by Marxist/Communist/Socialist/antifa elites is their enemy, deflecting attention from the major tech corporations and others algorithmically targeting their brainstems and manipulating their novelty bias in order to strip-mine their sanity and turn them into Eloi with debit cards.

And what of the opposition in what remains of America’s liberal democracy? The Democratic Party is ideally positioned to foster a major social reform movement against digital oligarchy in league with new corporatists, and the harm they do to the commons: It is national in scale, and has an organized presence downward in each and every state, town, and county. And yet in that regard it has done about as little as can be imagined under the circumstances. Why?

To be fair, Massachusetts Senator Elizabeth Warren has called repeatedly for the regulation and breakup of the major digitech corporations, and she and several others in the party, notably Connecticut Senator Chris Murphy, are volubly aghast at the politicized media consolidation antics we have been recently witnessing. Not very much active company have she and Senator Murphy attracted, however, from among their peers.6 They all know what’s happening, many of them, to one degree or another, but they say and do very little about it even amid their own state and local constituencies.

The Biden administration did throw the spear of antitrust litigation at the tech giants and other industries it viewed as overly consolidated, marking a break with the party’s friendly posture toward an industry long viewed as an ally. But Democrats are too feckless, divided, and unpopular to win the kind of mandate that might allow them to think bigger—only a 27 percent national approval rating at last measure, so in single digits in many if not most beyond-xburbs parts of the country. This is true not least because the social media snare has also trapped left-wing politics, fueling the woke ideologies and purity tests that have turned off so many Americans.

Will this change? Growing signs of discomfort and tumult over social media harms, and some glimmers of action in the culture and even in the courts, suggest it might. We may hope so, but as has been often been recited, hope is not a policy. In any event, we are bound to find out.

Adam Garfinkle is founding editor of The American Interest and a Senior Fellow at the Niskanen Center. Before founding the magazine in 2005, he served in 2003-05 as speechwriter to the Secretary of State.

Share

Sixty years. That’s how long it took Americans to organize a coherent political response to the predations of 19th- and early-20th century industrial capitalism. From the completion of the transcontinental railroad in 1869 to the codification of the New Deal in the late 1930s, the country lurched through Grange halls and union massacres, through Populist defeats and Progressive victories, through muckraking exposés and constitutional amendments. The response, when it finally came, was a genuine transformation: antitrust law, labor protections, financial regulation, the administrative state itself.

But the response to the latest economic revolution in the United States—the fourth, if you’re counting—marked by the internet, social media and AI, has so far not been as productive. We have disruption, polarization, a youth mental health crisis, and an economic concentration of wealth that would make the robber barons think maybe we’ve gone a little overboard. What we lack is a coherent political movement that offers a real alternative. Bans on phones in schools and Australia’s crackdown on social media for kids, along with a few lawsuits here and there, represent the most visible stabs at reform. But they feel less like the beginning of a genuine movement in the mold of the old trust-busting Progressives and more like the exertions of passengers trying to bail out a sinking ship with teacups.

The usual explanations for the tech industry’s dominance–regulatory capture, a polarized and gridlocked political environment, and the sheer complexity of digital technology–have merit. But they miss an even larger problem. The absence of a mass-scale, coordinated response isn’t necessarily a failure of imagination or political will; it may be a feature of the system itself.

To see this at work, start with the Big Tech business model and the concentrated wealth it generates. Cory Doctorow’s framework of “enshittification“ has become something of a meme, but underneath the coinage lies a sophisticated diagnosis. Platforms follow a predictable arc: At first, they’re good to users to achieve growth. Next, though, they exploit users to court business customers such as advertisers and publishers. Finally, with everyone locked in, they squeeze them all to satisfy shareholders. In 2008, Facebook promised not to spy on us; today, Facebook is a rent-extraction machine optimized for engagement metrics that correlate well with human misery.

But Doctorow’s framework needs an amendment. After the extraction phase, something else often happens. Not all that captured wealth goes into manipulating the regulatory environment to increase profits. Some is plowed back into the next wave of disruption, like Google Ventures, Microsoft’s $13 billion OpenAI bet, Amazon’s $8 billion Anthropic investment, and Meta’s pivot to AI (up to $65 billion in spending in 2025 alone). The enshittification of Platform A generates the venture capital financing for Platform B, thus creating a self-fueling flywheel.

The system doesn’t just extract and decay; it extracts, decays, and innovates, often simultaneously. That innovation makes it hard to dismiss the current system as one of mere corruption (although it certainly is corrupt). The corruption, for better or for worse, funds further innovation. And the innovation, in turn, resets the cycle: A new platform, a new promise, a new wave of users being treated well until they’re not.

This new dynamic represents a virtuous cycle for tech capital, but a vicious one for everyone else.

Add politics to the mix and the picture gets even grimmer. Tech wealth buys lobbying power. Lobbying power prevents regulation (or captures it, which amounts to the same thing). Unregulated platforms accelerate the extraction of wealth, which fuels another spin of the flywheel, each faster than the last. We’ve watched this play out with Section 230, with antitrust enforcement that arrives a decade too late, with privacy legislation that somehow never quite passes.

The flywheel, in other words, includes a political capture loop in its rotation. And unlike older forms of regulatory capture, it generates innovation while also extracting rents. The railroad barons could only lay so much new track. The platform barons keep building new products, finding new frontiers, dominating new markets.

There’s another, more profound difference. In the Gilded Age, economic and narrative power were separate.

The media industry wasn’t consolidated like it is today. When Upton Sinclair published The Jungle or when Ida Tarbell exposed Standard Oil, people like railroad magnate and financier Jay Gould might have been able to bribe or threaten individual editors, but they couldn’t suppress the story everywhere. Muckrakers like Sinclair and Tarbell found outlets. Populist newspapers reached farmers. Gould could buy legislators, manipulate shipping rates, and deploy strikebreakers against his workers, but he couldn’t control what farmers thought about him. The media environment was too diffuse and fragmented. The robber barons’ attempt at corruption simply couldn’t scale.

That’s no longer the case. Elon Musk owns a platform used by around 200 million people a day, including tens of millions who get their news there. He can, and does, algorithmically boost his own posts, suppress his critics through shadow-banning, and test messages in real time to see what resonates. He doesn’t need to bribe columnists. His engineers can just decide what those columnists’ audiences see or even if they see the columnists at all.

This poses severe challenges to one of the Gilded Age’s primary mechanisms for reform. Sinclair’s Jungle changed federal law because as powerful as they were, the lords of Packingtown couldn’t make the book disappear; they didn’t own the presses (for the most part). But platform barons like Musk and Zuckerberg today exert immense sway over the means of information production and distribution both, giving them outsized control over the media environment. The muckraking that forced Standard Oil to accept regulation doesn’t work when the monopolist owns the megaphone.

And yet, the political and social paralysis can’t be explained solely by an information stranglehold. People are genuinely angry. Scroll through any social media feed and you’ll find raw fury at tech companies, algorithms, and the attention economy devouring childhood. The energy and desire for a political movement seem to exist. So why doesn’t it gel?

Look at what the Progressive Era’s response actually required: Lots and lots of patient, face-to-face, grinding work. Meetings at settlement houses and union halls. Investigative journalism sustained over years, not news cycles. All of it depended on repeated in-person contact, trust-building, and the slow accretion of solidarity despite racial and social differences.

Social media, on the other hand, is optimized for precisely the opposite. Outrage gets engagement; nuance doesn’t. Identity politics fragment potential coalitions. Viral moments dissipate like flash paper before anyone can form durable organizations around issues. Herbert Croly’s Promise of American Life, which became a “manifesto of Progressive beliefs,” according to the New Republic, took nine years to write and publish. Today we have Substack and 280 character limits.

The irony is obvious. The technological tools we could use to build a resistance infrastructure are controlled by the very forces we need to resist. We’re forced to organize on digital terrains owned by the very forces we seek to rein in. And when things get too hot for the platform barons, they shift the terrain. By the time you’ve built a coalition around one grievance, the platform has already moved on, generating new harms and new distractions.

In short, the flywheel generates disruption faster than responses can form. The platforms shatter the coalitions that might demand their regulation. Each turn of the cycle produces more capital for the next wave of innovation that atomizes the public further.

The Progressive Era was 60 years of patient organizing. Try getting someone to watch a TikTok reel of more than 60 seconds.

The Progressives found success when the crisis of the Great Depression met decades of organizing and capacity building. But they had to go through hell to get there. The Populists lost. William Jennings Bryan lost not only while leading them, but twice more while leading just the Democrats. The labor movement was shot and beaten for 40 years before the Wagner Act. Perhaps we’re in the “failing repeatedly” phase. That would be historically normal.

But what if the flywheel’s velocity has changed the math? Maybe the cycle now spins fast enough to outrun patient preparation. Maybe the platforms now so thoroughly control the online equivalents of union halls, where organizing might happen, that we can never quite prepare the ground. If that’s the case, then the politics adequate to this moment may never arrive.

The missing response isn’t an accident or a failure of nerves. The system in need of reform is the same system that prevents reformers from organizing. The flywheel spins, the opposition scatters, and the politics we need keep slipping out of reach.

So what can we do? The historical precedent offers cold comfort, but it’s not entirely bereft. Union organizing is having a small renaissance in the form of workers at Starbucks and Amazon. The most effective movements will combine offline solidarity with online amplification. And local politics are still relatively unmediated. The platforms haven’t entirely colonized every organizing space

Capacity-building will matter when a crisis hits, even if we don’t know what an equivalent crisis might be. It could be a catastrophic platform failure, an AI incident, a tech-driven financial collapse. But that capacity-building during the losing phase still matters. The Progressives didn’t know the Depression was coming, but they were ready when it did.

Christopher Allbritton is the executive editor for The Media Copilot, covering AI adoption in journalism and newsroom transformation. He has covered conflicts in Iraq, Lebanon, and Pakistan, including the killing of Osama bin Laden as Reuters’s Pakistan bureau chief.

Share

Call it the information age, the fourth industrial revolution, the cognitive era. Call it whatever you like, but it seems clear we are in the midst of an economic transformation as profound as those that sent medieval serfs to the cities, yeoman farmers to conquer the Western frontier, huddled masses to the United States.

The birth of industrial capitalism in America’s Gilded Age created the outlines of our political, economic, and social order, from the modern presidency to the eight-hour day to the unionized family wage. Facing mass immigration into overflowing cities, dirty and dangerous factory work amid machines of unfathomable power, and behemoth corporations that fundamentally changed the dynamics of economic competition, Americans of the late 19th and early 20th centuries banded together in myriad political and social movements attempting to tame the beast, from settlement houses to temperance movements to Theodore Roosevelt’s Bull Moose Party. Whatever their excesses and flaws, these movements preserved capitalism’s essential core of creative destruction while drawing a protective membrane around the social fabric that it threatened.

Today, the transformations of capitalism work on our minds, not our bodies. Their power is no less profound, from the acceleration of polarization to the youth mental health crisis to the usurpation of our basic cognitive processes by AI. Yet there has been a strange passivity in the reaction. Nobody likes social media, but nobody logs off. Suggestions of regulation, restriction, or modification are met with shrugs. We are told technology’s dominance of our private lives, families, and workplaces is inevitable. Everyone else is doing it, and there is no way out. Why this pervasive sense of powerlessness? Digital media is acting on our politics in obvious ways, but our politics don’t seem to be acting much on digital media. What will it take to rouse a coordinated reaction—or has one perhaps just gotten started?

When reflecting on the social response to Gilded Age capitalism, it’s important to remember how long change took, and the many channels through which it flowed. The last era of capitalist reform stretched for roughly 60 years, from the 1869 completion of the transcontinental railroad, which supercharged economic growth, to the end of the 1930s, when Franklin Roosevelt’s New Deal codified in federal legislation worksite policies first pioneered on the state level. What makes it into the history books is landmark legislation and constitutional change. But these six decades brought a variety of interlocking responses responding to the ripple effects of industrial change in society and culture. Courts had to work out just who was responsible when a railroad took life or limb. Temperance crusaders had to coax women out of the home and into public life to agitate against the saloon, that great destroyer of families. Workers were shot and killed with regularity along the path to safe working conditions, from the Homestead Strike of 1892 to the 1932 Ford Hunger March. Varieties of unionism rose and collapsed. College-educated elites conducted studies, ran for office, and exposed corruption. Urban bosses organized immigrant voters and constructed patronage systems that doubled as a proto-welfare state.

Critically, as industrialization proceeded, its booms and busts began to create common cause between American farmers, urban immigrant workers, and their champions in the professional class. The central dividing line of mass politics — tariffs — had long split agrarian communities from manufacturing towns. But now both found they needed something new from politics. It’s as if today the rise of artificial intelligence threw blue and red states together into a new order that erased the diploma divide. Back then, it took crisis upon crisis, campaign upon campaign, until a new syntax of politics emerged in books like Herbert Croly’s Promise of American Life and Walter Lippmann’s Public Opinion, which both called for a new politics to meet a new era, providing a vital rallying call for political leadership.

It’s also important to remember how overwhelming this change was to those who lived through it. Railroads cut time and space to ribbons. At one time in our history, the vast majority of Americans were used to moving no faster than their feet could carry them — or at the speed of a horse, if they were lucky. It must have been profoundly disorienting — albeit thrilling — to watch the landscape blur through a train travelling in hours a distance that once took days. Standing before the belching iron horse, what American could believe they had the power to impact this stupendous triumph of technology, this iron god? Likewise, today it seems impossible to imagine that digital technologies could be reshaped by political or social pressure, given their ubiquity, their usefulness, their addictive qualities. The problem seems even more intractable given how digital mediums have become indispensable to the cultivation and maintenance of political power. If the changes digital technology has wrought are less tangibly obvious than the railroad, the steamship, the industrial factory, or the automobile, they are no less overwhelming when taken together.

Yet it is possible to glimpse the emerging contours of a comparable politics, starting with the stirring of diffuse social movements and reactions. Someday, we might look back to Jonathan Haidt’s The Anxious Generation and place it in the same category as Upton Sinclair’s The Jungle, the expose which forever changed American factories. The movement to ban phones in schools is one of the first organic social responses to the new world we have entered. From silent book clubs to the Reconnect movement on college campuses to the consortium of authors that sued Anthropic, the signs of resistance are everywhere. This transformation in culture is essential. Before the tobacco litigation of the 1990s, for example, came a widespread recognition that smoking was harmful, and a decline in social acceptance.

Slowly but surely, Americans are recognizing the naivete of our first gee-whiz reaction to the internet, which was imagined by both political parties as a new zone of freedom and celebrated with inane slogans like “information wants to be free.” But the problem is our laws and norms date from that era. The time has come for a rethinking based on what we now know — that data wants to be expensive, cyber surveillance is unavoidable, and the hours spent on social media have wreaked havoc in the real world.

Until the shooting of Charlie Kirk, there was no national political figure willing to take on this cause. But when the national spotlight swung to Utah Governor Spencer Cox, Americans heard the first serious case against social media from a political leader — one driven by tragedy, but that nonetheless offered a studied reflection on what technology has wrought. “‘Cancer’ probably isn’t a strong enough word,” Cox declared in the aftermath of Kirk’s death, going on to call social media algorithms “evil.” This religiously inflected language may sound jarring at first. But the passion it channels is exactly what powers effective political movements. Cox too has a target in mind: “these companies with their trillion dollar market caps … They’re hijacking our free will with these dopamine hits, same chemical reaction as fentanyl.” It’s a populist take that mirrors the diagnosis once aimed at the beef trust and Standard Oil.

What would a politics that squarely confronts the change wrought by the fourth industrial revolution look like? It might look like the Utah legislation that Cox championed, requiring robust age verification, that has been stalled by lawsuits. Cox’s reference to fentanyl suggests it could look like the opiod lawsuits, and indeed suits against social media and AI platforms for a range of harms are wending their way through the courts. It could look like legislation to make public the algorithms that shape platforms, or a requirement that makes consumers the owners of their data. National security concerns around TikTok suggest another dimension to the problem — both the recognition that adversaries are actively using these platforms to sow internal dissent, and the powerful forces that resisted legislation.

Charlie Kirk’s shooting may accelerate the politics of reform. Until recently, there has been little incentive for political elites to rein in digital media. Like Kirk, after all, many owe their careers to its successful use. But political violence — and the widening of targets beyond national-level politicians to state-level leaders, organizers, and activists — may change the calculus dramatically.

The tech companies have an edge — they control much of what we see, feel, and think. But the farmer who needed to pay exorbitant rates to the railroad seemed powerless too. So did the recent immigrant with nothing to offer but their labor. What made them strong was a shared sense of purpose, legitimate grievance, and political leaders willing to respond. The last time around, it took nearly half a century to marshal the social, political, and legislative forces to meaningfully alter the fallout from industrialization. Today, technology itself has sped up the clock, so it might be later than we think. But if capitalist upheavals are coming faster than ever before, maybe so can the response.

Share

Jennifer Burns is Edgar E. Robinson Professor of United States History at Stanford University and a research fellow at the Hoover Institution. She is the author of Goddess of the Market: Ayn Rand and the American Right and Milton Friedman: The Last Conservative.

Image: Lawrence strike, 1912 , Digital Public Library of America. Modified with ChatGPT.

A webinar on the demographic shifts that drove the 2024 election.

Check out the recent commentary from our panelists:

• Amanda Iovino, “This is the Fault Line That Defines the Election,” The New York Times.

• Thomas Edsall, “How Resilient Is the Emerging Trump Coalition?” The New York Times. (Quoting Matt Grossmann.)

• Patrick Ruffini on The Ezra Klein Show.

• Ruy Teixeira, “Trump Called His Win a ‘Historic Realignment’ of U.S. Politics. We Have Our Doubts,” The New York Times.

• Matt Grossmann, “The rise of the caps and gowns,” Hypertext.

The closer we get to Election Day, the more helpful it is to step away from horse-race coverage and last-minute political gambits and ask: How did we get to this point?

The most compelling answer contains a deep irony. We live in an era in which the elite that wields the bulk of power in our society has become broader and more diverse than ever before. And yet, that elite has seldom been more unpopular.

“Elite” is obviously a contested term, but in this issue of Hypertext we feature two critically important new books that use it roughly in the same way, to describe the nearly two-fifths of Americans 25 and older who hold at least a bachelor’s degree. Along with that education, most of them have acquired an enormous economic advantage — and a distinctive worldview.

All of this means that the much-discussed “diploma divide” defining elections in the U.S. (and other wealthy democracies) runs far deeper than the polling place. David Hopkins and Niskanen Senior Fellow Matt Grossmann explain the stakes in their new book Polarized by Degrees:

Culturally progressive technocracy, the governance of society by socially liberal and well-educated experts, is winning a long-term battle, reshaping the governmental, business, and nonprofit sectors – but not without stimulating a major backlash that has redefined conservative politics. 

With Harris and Trump at an apparent dead heat, pundits across the political spectrum are puzzling at why neither party seems capable of seizing the obviously available middle ground in American politics (see recent pieces by Yasha Mounk, David Brooks, and Liam Kerr — all citing the work of our friends Ruy Teixeira and Yuval Levin). Both parties certainly have room to moderate and, as we recently highlighted, to reassert themselves as vital organizations: A functioning, well-staffed, enduring party is the precondition for connecting with voters — and for desanctifying the commentariat “priesthoods” that, Brooks argues, now dominate the GOP and the Democrats.

But it may also be true that overcoming the educational break among Americans is harder than moderating, managing, and messaging. The diploma divide has become a gulf.

Meanwhile, a broader elite itself is awfully difficult to unify, particularly when its ranks are filled with people who make their living manipulating abstractions and who believe that words and symbols are the most vital political goods, as Musa al-Gharbi explains in a terrific new critique, We Have Never Been Woke. 

These “symbolic capitalists” virtually all hold college degrees and as such have far better prospects than less educated Americans, al-Gharbi explains. They nonetheless will include many frustrated aspirants who cannot land the best jobs for which they are qualified. That frustration leads to a status competition among the elite’s relative winners and losers — a competition that takes the shape of claiming that one’s rivals are not living up to their professional obligation to advocate for the less fortunate. The irony, of course, is that the less fortunate generally could care less about these disputes — and more often than not, al-Gharbi notes, symbolic capitalists wage these fights while exploiting those people’s labor.

The “elite” in these stories is broad indeed, lumping adjunct professors who are barely scratching out a living with billionaires who are pouring money into political influence. Even Peter Turchin, the theorist of “elite overproduction” who predicted our age of instability, focuses his attention on people with a realistic shot at the highest positions of economic and political influence in society.

But Grossmann and Hopkins and al-Gharbi provide a compelling case for taking this wider angle. The resentments that billionaire political donors are able to exploit are not purely ginned up by a devious Republican Party. Nor is there a conspiracy of Ivy League intellectuals forcing “wokeness” down Americans’ throats.

Meanwhile, the differences between people who are highly educated and people who are not extend beyond cultural attitudes, into life expectancy, well-being, and critically, social status. Brink Lindsey provides a window into this deeper decline in social position in his essay, “The declining status and leverage of ordinary people.”

The upshot is that the transformations in our culture run deeper than politics — and so do our problems. As Lindsey has put it:

When people were telling pollsters that they no longer trusted government, what they had lost confidence in — what they were in the process of giving up on — were established institutions and governing elites. Rational-legal authority was gradually disintegrating, and people were reverting to charisma.

This collapse of legitimacy is the product of deep economic trends that have even deeper social ramifications, and with which we have not yet reckoned. Even if we avert the demagogue’s rule this time, our troubles will not be over until we do.

Read the excerpts and essays:

Grossmann and Hopkins: The rise of the caps and gowns

Grossmann and Hopkins: How conservatives lost the institutions

al-Gharbi: How elite competition turns into culture war

Lindsey: The declining leverage and status of ordinary people

Steven Teles: The other diploma divide

Geoff Kabaservice: Trump's triumph over the establishment

Please read, share widely, let us know what you think, and as always, pitch us a response essay if you are so moved.

David Dagan (@DavidDagan, @daviddagan.bsky.social) is Director of Editorial and Academic Affairs at the Niskanen Center.

Share

Subscribe now

Future histories of our populist era may well begin with the 2008 financial crisis and the revolts it touched off. On the left, the short-lived Occupy Wall Street movement gave voice to widespread anger about economic inequality and maximalist demands to rein it in — largely from denizens of the professional classes, who now drive Democratic politics. On the right, rage over Republican support for the bank bailouts culminated in the Tea Party, which then drove the GOP’s mutation into Trumpism.

On both sides, Wall Street titans’ collapse and subsequent frantic rescue seemed like a metaphor for capitalist malaise — a system run increasingly on shell games and insider protection over innovation and market discipline, with ordinary people left behind. A political and intellectual reckoning has since led some to declare the end of “neoliberalism.” 

But for the sector that kicked it all into motion, relatively little has changed.

That is the grim reality Anat Admati and Martin Hellwig laid out in their epic explainer, “The Banker’s New Clothes,” written in the wake of the 2008 financial crisis and now updated for the post-Trump era of regulation. In the book, the authors persuasively argue that the surest way to prevent a repeat of the 2008 disaster is to require that banks, like most other corporations, fund their operations with much less debt and much more equity investment. 

This volume of Hypertext, a forum on the book, leads off with “Capital confusion,” an excerpt that makes this case. You’ll also read:

• Former Kansas City Fed President and FDIC Vice Chair Thomas Hoenig on “The myths that fuel the megabanks”

• Admati and Hellwig on “Bailouts forever”

• Sociologist and Niskanen fellow Colleen Eren on “Giving power to truth”

• My essay, “It’s time to challenge Wall Street from the supply side.”

The debate is happening now

The updated book is a timely intervention: Even now, banks are lobbying fiercely against a proposal, developed in the wake of last year’s Silicon Valley Bank debacle, that would indirectly nudge up equity requirements for the very largest institutions—those with assets over $100 billion. Regulators are reportedly preparing to back down even though most banks already have enough equity to comply, anyway.

There is a simple reason why banks don’t want to raise more money by reinvesting earnings or selling new shares of stock: They can borrow at a discount, thanks to the continued assumption in the markets that the government will bail them out again if things go sideways.

As always, former Kansas City Fed Chair Thomas Hoenig explains, they tout the myth that requiring more equity would inevitably stifle lending: 

“Preceding the Great Financial Crisis of 2008, I was repeatedly told by those in the banking industry that requiring banks to rely on investor capital would deprive small businesses and individuals of much much-needed loans necessary to grow their businesses. These arguments carried the day, as many of the very largest banks in the country were allowed to fund themselves with as little as 3 percent of investor money for every dollar of assets on their balance sheets. When the crisis erupted … the low-capital approach hardly left those small businesses and individuals better off: Loan volume declined most among the largest, less well-capitalized banks.” 

But while equity ratios in the single digits are unsustainable economically, they have proven themselves remarkably stable politically. As Admati and Hellwig explain throughout the book, politicians are cowed by the types of warnings that Hoenig recounts, or motivated by parochial concerns such as campaign donations, financing for pet projects, or backing their big banks against those of other nations.

Sociologist and Niskanen Center fellow Colleen Eren largely accepts this analysis but emphasizes that reckoning with how the government deploys finance for popular purposes requires us to push deeper. 

“The U.S. government has been implicated heavily in many of the institutional changes in banking that led to the crisis. And, importantly, it has done so not out of some kind of nefarious plan to ensconce an aristocracy and immiserate the masses, but to the contrary, to expand the vaunted American dream of a middle-class lifestyle to a larger subset of the populace—without increasing government debt … It has been countered that to the extent lax equity requirements promote these goals, there are much more efficient and less-risky ways to go about it, such as direct subsidies. But as Niskanen Center Senior Fellow Monica Prasad has argued, over the decades, ‘Financialization became the political path of least resistance for addressing citizens’ social needs.’”

It’s also unclear, Eren notes, how to mobilize a constituency that will break up these political alliances and demand fundamental change. 

In my essay, I argue that the emerging “abundance” movement is equipped to tackle this nasty problem. It’s an opportunity for a community seeking to revitalize the supply side of the economy and build a more capable state to address distortions (or market failures) in capital deployment and improve an impossibly complex regulatory regime by lowering the stakes of failure. Meanwhile, core commitments of the movement, such as boosting housing supply, could relieve some of the coalitional pressure brought on by the tradeoffs.

We hope you enjoy the essays, share them widely, and let us know your thoughts on this vital debate. On democracy’s balance sheet, economic failure is paid for in political legitimacy. We can’t afford another big write-off.

Share

Subscribe now

Image: Generated with Microsoft Image Creator AI

The third issue of Hypertext, a collaboration with our friends at Vital City magazine, features a star-studded cast of authors contemplating the prospects for “evidence-based policymaking” over 19 (count ‘em) essays.

The notion that policy decisions can be guided by sober, scientific analysis of “what works” is widely attractive, and indeed, it’s an important component of the Niskanen Center’s advocacy.

But what if the best evidence suggests that virtually nothing policymakers are trying moves the needle?

That is the central claim of a new and hotly debated article on criminal justice policy by Megan Stevenson, an economist at the University of Virginia’s law school: “Cause, effect, and the structure of the social world.” The article, which is well worth reading in its entirety, surveys the literature on randomized controlled trials (RCTs) of criminal-justice interventions. Stevenson argues that the vast majority of interventions evaluated with this “gold-standard” method “have little to no lasting effect” and this tells us something important about the reach of policy. “When it comes to the type of limited-scope interventions that lend themselves to high-quality evaluation,” she writes, “social change is hard to engineer.”

Is it really that bad? Is the whole evidence-based model misguided? Must we act more boldly — and can we? Are policymakers back to navigating by ideology and gut?

We and our partners at Vital City asked leading lights from academia, philanthropy, advocacy, government, and journalism to weigh in — and they gave us a lot to chew on. To highlight just a few of our terrific contributions:

What can we know and how can we know it? John MacDonald and Kerri Raissian warn against simply reading the RCT literature as a counsel of despair. Meanwhile, John Maki argues that public policy has an irreducibly interpersonal dimension that is almost impossible to pick up on with quantitative studies. Jonathan Rauch agrees: “RCTs have their place — and we should keep them in it.” 

Incrementalism and failure are the way of the world. Aaron Chalfin argues that RCTs accurately reflect the difficulty of changing the world, but this hardly means progress is impossible. Matt Grossmann points out that progressives and conservatives interpret the failures of “evidence-based policy” in opposing directions, and argues incrementalism is a product of political constraints that cannot be wished away. And the alternative, Philip J. Cook and Jens Ludwig add, is recklessness: “You only live once, but for how long?”

Meanwhile, research and policy do get better.  John Arnold notes that the bar on research quality is rising, and useful findings are emerging as a result. Jen Doleac explains the key improvements we have made in recent years and calls for a “science of scaling” for successful interventions. Jeffrey Liebman takes a more pessimistic view of the literature but calls for more experimentation, not less, and counsels against mandating “evidence-based policy” in ways that lock in programs with weak support.

There are concrete takeaways for policymakers. Alex Tabarrok argues that changing people’s preferences rarely works - but changing their incentives and constraints can. Jen Pahlka notes that if policy change requires ongoing, laborious cultivation to succeed, the same is true of the bureaucracies pursuing that change. Richard Hahn envisions a bureaucracy that wrestles away the academic monopoly on rigorous evaluation to learn faster.  

Finally, Tracy Palandjian and Jake Segal offer some valuable perspective: “We see good evaluations … as momentary glimpses of underlying truths about the world, the tips of icebergs poking out over a dark sea. They shift under our feet the moment we step — nothing is ever ‘proven’ — but it’s better than jumping, unseeing, into the water.”

More: Since this post was written, we got another batch of terrific essays from Chloe Gibbs, Sherry Glied, Marc Levin, Candice Jones, and Anna Harvey.

Please visit Hypertext today to start browsing the entire issue, and look for individual essays in your inbox every few days over the coming weeks.

As always, please spread the word about Hypertext and help us broaden these critical conversations. And keep in mind that we publish response essays — if you know someone who wants to keep the debate going, have them reach out. Finally, don’t forget to sign up for updates from our wonderful partners at Vital City, who are building a hub for thinkers and doers to keep making policy better.

Share

Subscribe now