Lina Khan: The Regulator Who Tried to Put a Leash on Big AI

In 2017, a 28-year-old law student published a paper in the Yale Law Journal that would make her the most feared woman in Silicon Valley. The paper was “Amazon’s Antitrust Paradox,” and it argued that the existing framework for evaluating monopolies — focused on whether consumers pay higher prices — was fundamentally broken for the digital age.

Four years later, Lina Khan became the youngest Chair of the Federal Trade Commission in history. By 2026, her regulatory philosophy has been tested, challenged, and debated more intensely than any FTC Chair’s in decades — particularly when it comes to artificial intelligence.

The Thesis That Changed Everything

Khan’s core argument, distilled: traditional antitrust looks at whether consumers are harmed by higher prices. But in digital markets, the product is often free (Google Search, social media) or below cost (Amazon goods). The harm isn’t price-based — it’s structural. Companies use their platform power to crush competitors, acquire potential rivals, and lock users into ecosystems they can’t leave.

Applied to AI, this thesis becomes explosive. Consider:

• OpenAI raised $40 billion from Microsoft, which also provides the Azure infrastructure OpenAI runs on. Is this an investment or an acquisition by another name?
• Google controls the search engine (data collection), the cloud platform (training infrastructure), the browser (distribution), and the mobile OS (deployment). Can any AI startup compete with this integrated stack?
• Meta releases open-source models that smaller companies can’t afford to train, simultaneously claiming to support open AI while making it impossible for funded competitors to justify their own model training costs.

Khan saw these dynamics and acted.

The FTC’s AI Agenda Under Khan

Investigation 1: AI Training Data and Copyright

In 2023-2024, the FTC launched investigations into how AI companies collect training data. The core questions:

• Did AI companies scrape copyrighted content without permission to train models?
• Did they violate their own terms of service by using user-generated content for training?
• Did they adequately disclose how user data would be used in AI systems?

The investigations targeted OpenAI, Google, Meta, and Stability AI. The FTC issued 6(b) orders — compulsory information requests — demanding detailed documentation of training data sources, licensing agreements, and data processing practices.

What came of it: Several consent orders requiring companies to disclose training data practices in their terms of service. No major fines, but the investigations forced transparency that hadn’t existed before. Meta notably added explicit “AI training” disclosures to Instagram and Facebook terms of service as a direct result.

Investigation 2: AI Deals and Acquisitions

Khan’s FTC scrutinized the financial relationships between AI companies and Big Tech more aggressively than any previous commission:

• The Microsoft-OpenAI relationship was investigated as a potential de facto acquisition. The FTC’s position: Microsoft’s $13 billion investment, exclusive cloud hosting agreement, and commercial product integration with OpenAI technology effectively gave Microsoft control without triggering merger review requirements.

• The Amazon-Anthropic investment ($4 billion) received similar scrutiny. The FTC questioned whether Amazon’s investment, combined with a cloud hosting agreement, constituted an acquisition that should have been reviewed under Hart-Scott-Rodino.

• Google’s acquisition of DeepMind (2014) was retroactively examined as a case study in how Big Tech uses acquisitions to prevent AI competitors from emerging.

What came of it: The FTC issued new guidelines for AI-related investments, requiring companies to notify the commission of “non-traditional acquisition structures” (investments + exclusive partnerships) that function as acquisitions. This was the most significant structural change to merger review in years.

Investigation 3: AI Consumer Protection

The FTC took enforcement action against companies making deceptive AI claims:

• Fake reviews generated by AI: The FTC fined companies that used AI to generate fake product reviews on Amazon and other platforms. The agency established that AI-generated reviews are deceptive advertising, regardless of whether they’re disclosed as AI-generated.

• AI impersonation: New rules requiring disclosure when consumers are interacting with AI systems, not humans. This targeted customer service chatbots, AI-generated sales calls, and AI avatars in video content.

• “AI washing”: Companies that claimed products were “AI-powered” when they used simple rule-based systems or basic algorithms faced enforcement action. The FTC established standards for what constitutes a legitimate “AI” claim.

The Criticism

Khan’s approach generated fierce opposition from multiple directions.

From the Industry

AI companies argued that aggressive regulation would push AI development overseas. The talking points were predictable:

• “China doesn’t regulate AI this way”
• “Innovation requires regulatory freedom”
• “The FTC doesn’t understand the technology”

These arguments have some merit. The Microsoft-OpenAI investigation, regardless of its outcome, created uncertainty that affected fundraising for AI startups. Venture capitalists reported that potential investors in AI companies asked whether FTC scrutiny was a risk factor — a direct chilling effect on investment.

From Progressives

Surprisingly, Khan also faced criticism from the left. Some argued she wasn’t aggressive enough — that investigations and consent orders were weak responses to genuine monopoly power. The academic critique: Khan identified the problem correctly but the FTC’s enforcement tools, designed for a pre-digital era, were insufficient to address it.

From the Courts

Khan’s FTC lost several major cases in court. Judges appointed across political eras pushed back on the FTC’s expansive interpretation of its authority, particularly regarding:

• Whether investment partnerships (as opposed to traditional acquisitions) fall under FTC review authority
• Whether the FTC can establish new regulatory frameworks without Congressional action
• Whether the “unfairness” doctrine can be applied to AI practices that have no clear consumer harm

The Legacy Question

Khan’s tenure raises a fundamental question about AI governance: Is regulation-by-enforcement the right approach, or does AI need purpose-built legislation?

The enforcement approach (Khan’s method):

• Pros: Fast, doesn’t require Congressional action, adapts to specific situations
• Cons: Inconsistent, depends on which cases are brought, reversed by future administrations

The legislative approach (EU AI Act model):

• Pros: Comprehensive, predictable, democratically legitimated
• Cons: Slow, potentially outdated before implementation, one-size-fits-all

Khan chose enforcement because legislation was politically impossible — Congress couldn’t agree on AI regulation any more than it could agree on anything else. Whether this was pragmatic or inadequate depends on your perspective.

The Impact on AI Development

Regardless of whether you agree with Khan’s approach, her FTC has measurably changed the AI industry:

1. Transparency improved. AI companies now disclose training data practices, model capabilities, and limitations more thoroughly than before. This was driven directly by FTC investigations and the threat of enforcement.

2. Investment structures changed. The post-Khan investment landscape includes more careful structuring to avoid triggering FTC scrutiny. Whether this protects competition or just teaches companies to better disguise acquisitions is debatable.

3. Consumer protections exist. AI disclosure requirements, fake review enforcement, and AI washing rules are tangible consumer protections that didn’t exist before Khan’s tenure.

4. The debate is framed. Khan’s enforcement actions established the intellectual framework for AI regulation. Even critics who disagree with her methods now debate AI monopoly, data rights, and consumer protection using the terms she defined.

Who Is Lina Khan the Person?

Behind the regulatory firebrand is a person whose career trajectory defies Washington norms. Born in London to Pakistani immigrant parents, raised in New York, Khan attended Williams College and Yale Law School. She was a journalist before she was a lawyer — writing for publications about platform monopoly power.

Colleagues describe her as intense, prepared, and surprisingly good-humored given the constant attacks from industry lobbyists and political opponents. She reportedly reads every major filing personally and has been known to call staff at unusual hours to discuss case strategy.

Khan’s personal philosophy, expressed in interviews: “Markets should serve people. When they don’t, the government has an obligation to act.” Applied to AI: “The question isn’t whether AI should exist. It’s whether a handful of companies should control the infrastructure of intelligence.”

What Comes Next

Khan’s future is uncertain — FTC chairs serve at the pleasure of the president, and political winds shift. But the framework she established — applying antitrust and consumer protection principles to AI companies — will outlast her tenure.

The AI industry got its first serious regulator. Whether she was right, wrong, or somewhere in between, Lina Khan ensured that the growth of artificial intelligence wouldn’t happen without someone asking uncomfortable questions about power, competition, and fairness.

In an industry full of people building the future, she was the person asking: “Who gets to decide what that future looks like?” That question isn’t going away.