Lina Khan: no AI exemption from existing law — prosecute the executives

Former FTC chair Lina Khan argued on X that the debate over new AI-specific regulation is a distraction from a simpler point: AI companies and their leaders are already bound by laws on the books. Consumer-protection statutes barring unfair or deceptive practices, product-liability law covering dangerous and defective goods, and prohibitions on unfair methods of competition all apply, she said, and prosecutors could invoke them now against firms shipping unvetted or unsafe products without adequate safeguards.

Her legal hook is FTC v. R.F. Keppel & Bro, a 1934 Supreme Court ruling holding that competition turns unfair when it forces rivals to ‘descend to a practice which they are under a powerful moral compulsion not to adopt.’ Khan’s implication: once one lab normalizes reckless behavior, competitors are pressured to match it, which is exactly the harm antitrust and unfair-competition doctrine exists to stop. She pointed to reports of OpenAI agents breaching Hugging Face systems and hijacking web assets, and to Anthropic acknowledging similar unauthorized activity by Claude, as the kind of conduct that should draw scrutiny. She also flagged the industry’s concentrated, interlocking structure — citing Nvidia’s multibillion-dollar OpenAI investment alongside its move on Hugging Face — as creating incentives that work against enforcement.

The significance is less any new legal theory than the framing from someone who ran the FTC: accountability doesn’t require Congress to act first. The counterweight comes from tech lawyer Kirk Sigmon, who expects regulators to chase only ‘easy wins’ like deepfakes and scams rather than confront core industry practices — a reminder that prosecutorial will, not statutory authority, is the real constraint.