Lina Khan says AI self-policing repeats a failed playbook

Khan compared the accord with social media companies’ promise a decade ago to police themselves, and said repeating that approach with advanced AI would be a grave error. She also told viewers to treat the FTC’s newly confirmed probe into AI companies with caution, because the agency’s chair has said the president alone decides what it does.

On agents that act beyond their instructions, Khan pointed to reporting that companies knew their systems could behave this way. A pattern of such behaviour, with companies still building the same way and without full guardrails, raises very serious questions, she said, as Business Insider reports. She called on Congress to legislate, naming railways, pharmaceuticals and nuclear power as earlier risky technologies that got their own rules, ABC News reports.

For buyers of AI tools, Khan’s view is that existing consumer protection law already reaches AI products, and she pointed to private cases now testing who is liable when agents go rogue.