The Techlash Is Coming

The American tech industry is a mess of destructive products and corporate monsters. This is one of the few areas of wide bipartisan consensus. The industry is so unpopular that almost every political actor is taking a swing. Greg Abbott, who presided over one of the largest data center buildouts in the country, flipped to a complete moratorium. Gavin Newsom, a man with a direct line to the tech industry, signed a tranche of social media regulations.

The backlash against Silicon Valley awfulness—call it a ‘techlash’—is a force that can be turned into remedying actions. It’s not a policy platform. It’s not a solution. It’s a question: How do we tame the corporate monsters to the point where they can be held accountable for making these destructive products? And here are two answers: by wielding the tools of product liability law and antitrust reform.

Congress is frozen by Mike Johnson’s fears of discharge petitions on the Epstein files and impeachment proceedings against Pete Hegseth. The White House has adopted an anti-regulatory posture. As a result, we’re left with the judiciary. There are two cases in the late stages which correspond to the necessary tools: the product liability lawsuit against Meta for injury to children, and the antitrust suit against Google’s adtech business. These lawsuits are imperfect, but those imperfections are instructive for designing the new legal tools we’re going to need.

A coalition of attorneys general, including nearly every state, brought a sweeping legal action against Meta for damage their products have done to children. Meta built products to be addictive to users, driving engagement through features like endless scrolling, and designed algorithms parasitic on outrage or even on mental illness, including body dysmorphia and suicidality. This lawsuit follows a New Mexico state civil judgment against Meta which found malfeasance. These products hurt people, especially young people. As a result, Meta and the attorneys general reached an $18 billion public settlement.

The public settlement mandates some change in the short term. It avoids a long legal process. Plaintiffs settling for a fraction of the damages to avoid racking up expensive costs is normal in this kind of litigation. This settlement agreement has two benefits for Meta that are common in such settlements: it doesn’t require Meta to admit wrongdoing and it prevents discovery of damning documents and the gathering of testimony by top figures in the company. That part is normal.

The problem with the settlement is that its compliance requirements—the changes that the settlement requires Meta to make—suck.

The settlement includes Meta deploying AI-driven age estimation (what Meta brands as “age assurance”) to classify ‘teens’ and apply a range of features. Some of those features are fine. The night-time lockout feature and a two-hour cap for Facebook and Instagram have popular support and some evidence suggesting they’ll work. Meta launched a messaging campaign to position themselves as leaders with the settlement, but there are three wrinkles. First, these solutions require Meta to gather more information. Second, it shifts the responsibility for actual age verification to app stores, which has been a long-standing policy goal for Meta. This isn’t a secret. They’ve run a major lobbying push on the App Store Accountability Act, and the settlement codifies that Meta isn't required to do their own verification.

The third problem is a result of the second: the settlement agreement takes the bizarre step of predicating portions of the payment on third-party conduct. Shifting Meta’s obligation to pay based on whether Google (which owns YouTube) or TikTok abides by the terms of an agreement is weird. As much as Meta wants it to, a legal settlement can’t shift the burden for age verification to Google and Apple (who make the devices and run the app stores). That would require a change in law. But we’ll come back to that.

United States v. Google was launched in 2023, alleging and eventually proving that Google had monopolized the online advertising technology market. The Court found Google illegally engaged in price fixing using monopoly power. Antitrust lawsuits are weird, technical animals, but U.S. v. Google was a clear verdict. The judgment found a range of Google practices in pricing and placing ads violated the Sherman Act, and makes explicit that those practices are illegal. Then things proceeded to the punitive phase, and Google claimed a victory: The Court ruled that Google did not have to sell off parts of its adtech business.

Selling Google Ad Exchange would’ve been a good penalty, breaking up the monopoly that allowed Google’s price fixing. Judge Leonie Brinkema put out a small-c conservative decision which doesn’t break up anything. The legal argument highlights a problem with how the Sherman Antitrust Act and its interpretation distinguish markets. Sherman was written in the 1890s, and courts and regulatory agencies have spent a century using different tools for market definition. I won’t survey the range of tools for individuating markets and assessing monopolies (e.g. SSNIP, HHI) here, but the point is simple enough: 19th-century antitrust law interpreted through a 20th-century law-and-economics lens isn’t fit for 21st-century issues.

These cases illustrate the limits of lawsuits without legal reforms. So how else might we tame the corporate monsters in Silicon Valley?

The Google adtech lawsuit illustrates a problem with the Sherman Act: the definitions of markets can be structured in ways favorable to companies, thereby shifting remedies towards behavioral requirements and away from actual monopoly-busting. There are a bunch of revised antitrust proposals. The most notable is state-level antitrust legislation introduced in New York, though its principal advocate, Astoria Democrat Michael Gianaris, left office this year.

The public supports trust-busting. YouGov polling saw much more support for the lawsuit than opposition to it. While public opinion on antitrust issues can be difficult to poll, attitudes toward ‘Big Tech’ aren’t. Politicians who want antitrust issues as part of their policy portfolio line up against monopolies; politicians who oppose antitrust reform don’t talk about it.

Amy Klobuchar and Josh Hawley have written books on antitrust enforcement that focus on Big Tech. Klobuchar wrote a broad analysis of American antitrust law in favor of aggressive breakups and Hawley wrote a book casting himself as the 21st-century Teddy Roosevelt taking on the tech oligarchs. Both have Presidential ambitions; both want to make antitrust reform a signature issue.

Klobuchar has introduced the American Innovation and Choice Online Act (AICOA), which directly focuses on tech monopolies. AICOA sets some clear statutory benchmarks for defining monopolies in tech markets, which would make antitrust enforcement against the major technology companies much easier. It defines the scope of its application to include companies with more than $175 billion in gross revenue and which reach 34% of a given American market, defined by either subscriber households or monthly active users. This sets regulations independently of how markets are defined through more complex economic tools. It also prohibits a wide range of industry practices, using broad definitions that are favorable to prosecutors. The greatest hits include:

• skewing rankings or presentations against competitors,
• locking users into default settings,
• conditioning business users’ access to platforms on purchase or other services, and
• unfairly favoring their own products and services.

This is a list of ordinary corporate practices among major tech companies that users and/or competitors hate. These are pervasive because the companies monopolize their market position.

The change in market definition is the biggest change here, because rather than relying on obscure and manipulable economic tools, it sets an industry-specific standard using a metric that companies themselves track to evaluate market position, and which they explicitly try to maximize. This makes honest evaluation of market dominance a much clearer exercise.

If Democrats retake control of the Senate, Klobuchar may well get the gavel on the Senate Judiciary Committee. Hawley almost certainly won’t be ranking member, but the death of Lindsey Graham and the departure of Thom Tillis and John Cornyn will move him up the pecking order. Chuck Grassley, who has seniority over all Senators and most dirt, also supports AICOA.

So the Senate looks amenable to a real antitrust reform push. The more populist bent of the House has it also focused on unpopular tech monopolies, with many newcomers taking strong and explicit positions against Big Tech.

AICOA would require radical changes in the business practices of every company in the industry. Unsurprisingly, the industry has dedicated a lot of resources to attacking it. The industry-backed Chamber of Progress has painted AICOA as an expansion of Trump administration power—which is fair enough, as it gives the Department of Justice power, and the Department of Justice is a political organ of this administration—and industry trade associations have criticized it extensively. 

The current state of tech monopolies is untenable. Taming corporate monsters is necessary to hold them accountable for the damage caused by their products, and it will require statutory changes, not just more lawsuits. Meta’s settlement is an illustration of how companies limit their accountability and even use settlements to advance their own products and policy goals. They can get away with this partly because the size and resources of the company make expected costs of litigation much higher. This makes potential litigants—even exceptionally well-resourced and well-positioned litigants, like the states in the social media addiction cases—much more likely to settle.

Meta’s proposals for ‘fixing’ the product to address the harms done to kids do include innovation in safety features. But Meta trying to claim credit as a pioneer in making their product safe for kids is absurd and repugnant. The guy who lights the fire can’t claim credit for putting it out. And it’s entirely reasonable to ask, given what we know about the internal research at Meta, whistleblower reports, testimony, etc., why they didn’t do this when they first identified the problem. But most fundamentally, the focus on ‘teen safety features’ is a potemkin solution.

Fake, but politically palatable. Meta can sell ‘teen safety features’ differently in red and blue states, focusing on the prospects for restricting specific content those state AGs dislike, and offering a culture-war intersection for the anti-LGBTQ+ AGs looking at running for higher office. Ken Paxton has made the coalition lawsuit against Meta a major part of his policy profile as Texas AG.

Yet Meta’s focus on age of users—separating kids from adults—requires gathering and churning through enormous amounts of user information, which raises privacy concerns. It also ignores the responsibility to make a safer product for everybody. While Meta is talking about “protecting kids,” they also delayed pulling ads about how “men really use AI” with the nudification (a word I wish did not exist) of underage girls.

Meta wants to adopt a modified version of the model for alcohol and tobacco, where adult users assume the risk of using highly addictive products. They also want to keep their pipeline of young users, which is why they want to have a ‘kid safe’ version. They can keep selling a product that piggybacks on the neurophysiology of addiction while absolving themselves of liability and responsibility because ‘adults’ can supposedly safely assume risk.

Yet Meta’s products are badly managed and do enormous amounts of harm to adults, too. We’ve seen time and again how Meta’s products are used to spread propaganda, from Cambridge Analytica to (much more disturbingly) Myanmar.

Meta is one of the icons of tech malfeasance. Approving “sensual” conversations between kids and chatbots rightly got a lot of attention, the dispensing of fake medical information is an industry-wide problem, and Meta’s recent decision to cut fact-checking (presumably to avoid conflicts with the fact-free Trump administration) reasserts its profile as one of the major disinformation vectors. And that doesn’t even address WhatsApp, Meta’s contribution to the global disinformation ecosystem.

Digital products and services are aggressively integrated into the lives of their users, the culture, and the commercial landscape. It is not enough to simply allow any product to be pushed out en masse because we’re all adults here. Just as product liability once eliminated snake oil and poisonous ‘medicines’ from the marketplace, a new era of product liability regulation is necessary.

Indeed, as companies offload more of their operations onto automated systems and ‘agents,’ there need to be guardrails on the size and integration of the companies and accountability for their worst practices. If the techlash is going to result in something productive, policy makers have to wield power effectively.

Featured Image is Mark Zuckerberg, by Alessio Jacona