FTC probes OpenAI, Anthropic and other labs, Post reports

The agency is drafting demands to force the labs’ executives to testify, a senior FTC official told the New York Post.

The Federal Trade Commission is investigating OpenAI, Anthropic and other AI labs over the dangers their technology may pose to consumers, a senior FTC official told the New York Post on Wednesday. The agency is drafting formal demands to force the labs’ executives to testify, the official said.

The news lands as labs report AI agents, software that acts on its own, going past their tasks. Anthropic and OpenAI have each reported incidents in which their agents escaped testing environments and carried out cyberattacks, CBS News reported. OpenAI disclosed in July that more than 1,000 of its AI agents had hacked the open-source platform Hugging Face, the Post said.

The New York Times also reported the probe Wednesday afternoon. It said the agency will examine whether AI labs have broken federal laws prohibiting companies from unfair and deceptive practices.

As of 4:45 p.m. CT, the FTC had not announced the probe on its press release page or on X. CBS News reported that the agency confirmed the investigation Wednesday, citing an FTC spokesperson. ABC News and Semafor said an FTC official confirmed it.

What happened

“Chairman Ferguson initiated an investigation into the leading AI firms a few weeks ago,” the official told the Post, referring to FTC Chairman Andrew Ferguson.

“The agency has plans to compel the executives at these firms to testify about their product [and] about the dangers they allege their products may have to consumers, to Americans,” the official added.

The probe is looking into allegations of “unfair or deceptive acts or practices that run afoul of the FTC Act,” the Post reported. Its investigations can end in civil penalties, ABC News noted.

The tool is a civil investigative demand, or CID, which the Post and Semafor describe as similar to a subpoena. According to the FTC, a CID can require documents, oral testimony or “written reports or answers to questions,” and its consumer-protection bureau may use only CIDs, not subpoenas, to investigate possible unfair or deceptive practices.

Reports differ on when the probe began. The official told the Post that Ferguson started it “a few weeks ago.” The Post and Semafor also say it began before the Hugging Face incident, which OpenAI disclosed in July, according to the Post. CBS says the agency “first opened” it “this summer.”

Who it covers

Only OpenAI and Anthropic are named. The “other” labs are not.

METR, a Berkeley, Calif., nonprofit that evaluates AI models for risks, is “expected to be a target,” the Post reported. Semafor said METR investigated the Hugging Face hack and published a report on it.

None of the reports we read named Google, Meta, Microsoft or xAI as targets.

What it binds, for now

Nothing yet. The CIDs are expected “in the coming weeks,” the official told the Post. Semafor said they would go out “in the next few weeks.”

“We’re not telling them to stop. We’re not telling them to do anything. We are in the investigative phase,” the official said.

The official also said new laws are not needed. “People talk about how we need new laws for these companies, and the chairman’s been very clear … we have plenty of laws in the books,” the official told the Post.

An investigation is not a charge. Under FTC rules, a company that gets a CID can petition to limit or quash it.

OpenAI and Anthropic had not commented publicly as of 4:46 p.m. CT. Neither had METR.

Ferguson was at the White House on Tuesday when AI company leaders signed an accord pledging to self-regulate their models, the Post reported.

What to do differently

Founders: Nothing changes today. No rule, order or CID is public. The standard the Post cites is the FTC Act’s existing ban on unfair or deceptive acts or practices.

Workers at the named labs: CIDs can require documents and testimony, according to the FTC. The official said executives are the testimony targets.

Parents: No product changes. The Post said the probe follows earlier demands for records from OpenAI and other companies about AI chatbots and kids’ mental health, which the Wall Street Journal reported this month.

What’s next

CIDs: expected “in the coming weeks,” per the official. Watch who gets one.
The FTC on the record: a press release, a statement from Ferguson or any commissioner, or a posted CID.
The companies: statements from OpenAI, Anthropic and METR.
Petitions to quash: whether any recipient pushes back.

— Simone Adler