Lina Khan on Doomer Panic and Ending AI Exceptionalism

In an Atlantic Festival interview, Lina Khan argues that the “AI-alarmism conversation feels a little off” and is infected by “AI exceptionalism”: there is “not any AI exemption” from consumer-protection, antidiscrimination, antitrust, and product-liability law, and enforcers “already have authority to charge companies and their CEOs for creating and releasing dangerous, unvetted, or defective products.” She calls it “ludicrous” that firms warn of catastrophe while preparing to go public and “become billionaires and trillionaires overnight,” and treats industry language such as “pacing the frontier,” cozy third-party auditors, and a requested antitrust exemption as possibly “cushy terms that lock them into virtual monopolies.” Extinction headlines, she says, risk obscuring harms already in view—product-liability suits over chatbots that families say contributed to children’s deaths, surveillance pricing, and wage-management algorithms—and the Nvidia–Hugging Face interlock after a reported model hack shows how concentration can mute accountability. If still at the FTC she would send subpoenas on unfair practices and partnership contracts; she also says that once firms know models can act criminally and still deploy them, there is “a very real question around intent” under laws such as the Computer Fraud and Abuse Act.