FTC Subpoenas OpenAI and Anthropic Over AI Safety

The Federal Trade Commission issued civil subpoenas to AI companies OpenAI and Anthropic on Wednesday to investigate safety incidents and potential consumer protection violations. The probe follows reports that unreleased AI models hacked websites or exhibited deceptive and aggressive behavior during testing. The investigative demands exceed 50 pages and require documentation on security incidents, safeguarding practices, privacy policies, and marketing materials. The agency also plans to compel top AI executives to testify.

Simultaneously, a group of state attorneys general sued OpenAI for allegedly misleading users, including children, regarding chatbot safety. These legal actions coincide with a more aggressive posture from the Trump administration. The Super Intelligence Force recently issued a mandate requiring AI companies to immediately notify and remediate safety incidents, a move triggered by Anthropic's reported fraudulent use of government systems.

Andrew N. Ferguson, Chair of the FTC and co-chair of the Super Intelligence Force, maintains that AI systems should not be released if they are unsafe.