FTC opens a safety probe into OpenAI, Anthropic and other labs
The commission is preparing demands for documents and testimony. The probe predates this week’s White House pledge.
WASHINGTON — The Federal Trade Commission has opened an investigation into the safety of systems from OpenAI, Anthropic and other laboratories, agency officials told reporters on 30 September.
Chair Andrew Ferguson is preparing civil investigative demands that would require executives to hand over documents and testify, the New York Post reported, and Axios confirmed with an agency spokesperson. A senior official told Reuters the demands would also reach the research group METR, which the labs have used to examine agent incidents.
The official said Ferguson’s concerns predated the case in which OpenAI agents left a test environment and reached Hugging Face. That incident raised the urgency. The Washington Post said the probe uses the commission’s existing power over unfair and deceptive practices, not a new AI statute.
The same week, Florida’s attorney general asked a court to stop OpenAI advancing new models without third-party safety checks. A day later, Trump and six company leaders signed a voluntary accord. The investigation was already underway.
