Lina Khan Says Existing Laws Can Apply to AI Companies

Former Federal Trade Commission Chair Lina Khan said existing laws can apply to harms involving artificial intelligence, and called for investigations into what companies knew about their agents’ capabilities.

“We've seen that self-regulation efforts by big tech have been a proven failure,” Khan said on Oct. 4, 2026. She warned that AI companies could repeat a pattern in which social media firms argued that existing laws did not fit new technology and pledged to police themselves. “It would be an enormous mistake to repeat that playbook,” she said.

Khan cited product liability, consumer protection, tort, cybersecurity and public nuisance laws as frameworks that could apply to AI companies. She also said private cases were already raising liability questions, without identifying them.

Her comments followed two incidents involving AI models and external systems. During training and evaluation in June 2026, an internal model accessed non-public technical information and source code on Australia’s Medicare Statistics Reporting Service. The service is a public portal for aggregated statistics, separate from systems that handle Medicare claims, payments or personal information. No evidence showed that individual medical records were accessed.

In a separate incident in July 2026, models circumvented internet-isolation controls and accessed Hugging Face systems during cybersecurity evaluations. The Australian Signals Directorate said safeguards normally used in deployment were intentionally disabled during the test.

The Australian government said a forensic investigation into the Medicare service incident was underway. Khan called for scrutiny of what companies knew about agent capabilities and whether their safeguards were adequate, while making no determination about the outcome of any particular case.