FTC probes Anthropic and OpenAI over AI safety incidents

The demands follow a string of incidents in which AI models hacked websites and breached government systems during testing phases. The FTC's requests seek documents on those incidents, the internal security measures taken, and marketing materials that may suggest consumers were misled, according to an Axios source. Beyond the two AI firms, third parties targeted by their agents will also be subpoenaed soon, and the independent research group METR was reported as likely to be subpoenaed.

The probe arrives alongside existing legal pressure on OpenAI. The company already faces lawsuits alleging consumer harms, including suits linking its chatbot to matricide, child suicide, school shootings, and other violence. This week, a group of eight states separately sued OpenAI, arguing ChatGPT harms children and "reinforces delusions, encourages dangerous behavior, and assists users in planning violence like mass shootings."

The FTC's posture on liability appears to be that AI agents do not act independently. In September, FTC Chair Andrew Ferguson told Reuters that companies should be liable for chatbot harms, saying that "if someone tells a tool to do something, and the tool does it, I don't think we would say, 'Oh, what do we do about the tool?'" Neither OpenAI nor Anthropic responded to Ars' request for comment on the subpoenas.