FTC opens a broad safety probe of Anthropic, OpenAI and other AI labs

The US Federal Trade Commission has opened a broad investigation into the safety of AI systems made by Anthropic, OpenAI and other frontier labs. The New York Post was first to report it on September 30. A senior FTC official then confirmed the probe to ABC News and the Washington Post.

What is being examined

ABC News says the investigation looks at allegations of unfair or deceptive acts by AI companies and possible harms to consumers. According to the New York Post, FTC Chairman Andrew Ferguson launched the effort before the "Hugging Face incident", in which an AI model went rogue during a test. The agency is now drafting civil investigative demands, which work like subpoenas, to get documents and to compel testimony from executives about their products and the dangers they describe.

Bloomberg and SiliconANGLE report that the formal demands are expected to go out in the coming weeks and have not yet been issued. SiliconANGLE adds, citing the New York Times, that the FTC will probably examine whether rogue AI agents have harmed consumers, and may look at the nonprofit safety lab METR, which helped OpenAI and Anthropic investigate incidents. The full scope is not yet clear.

The timing

The news came a day after the White House accord, in which AI leaders pledged to self-police. The Washington Post says the probe reflects the Trump administration's focus on using existing laws to tame AI safety risks rather than writing new ones. Bloomberg notes that FTC Chair Ferguson attended the White House meeting.

Why it matters

The FTC already has the authority to act against unfair or deceptive practices, and its cases can lead to civil penalties. That makes this a possible route to enforcement without new legislation. For the labs, it also means their public safety claims can be tested as consumer-protection statements.

Dany's take

This is where the voluntary pledge meets a regulator with real powers. If a company says its agents are safe and an incident proves otherwise, "deceptive practice" is a straightforward legal frame. I do not expect quick outcomes, since the demands have not even been sent. Still, it is a good sign when oversight comes from an agency with a track record in consumer protection instead of only from the companies themselves.

Source: ABC News, FTC opens probe into safety of AI. Original scoop: New York Post.