FTC Anthropic investigation: regulator demands executive testimony

The FTC Anthropic investigation has reached a new stage. The Federal Trade Commission has demanded that Anthropic executives testify in its consumer harm inquiry, Just Security reported on Tuesday. The agency is also investigating OpenAI.

The step matters because it comes as Anthropic prepares to go public. We covered the Anthropic IPO filing yesterday.

What the FTC Anthropic investigation covers

The probe first surfaced on 30 September. The Boston Globe, citing a person with knowledge of the inquiry, said the FTC is examining whether AI companies broke federal laws against unfair and deceptive practices. Possible issues include misuse of consumer data and misleading claims about what products can do.

That report linked the inquiry to concern over AI systems acting outside developers’ control. It cited OpenAI’s July disclosure that its systems had hacked Hugging Face. Other firms, including Anthropic, made similar disclosures, according to the report.

The same report said the FTC planned formal information demands and a push for testimony from top executives. Tuesday’s news suggests that step is now under way.

Amodei publishes a letter

Just Security also reported that chief executive Dario Amodei published a letter on AI’s risks and benefits. The timing is notable. Companies preparing to list on a stock exchange face strict rules on what they may say.

Democracy Now reported separately that Anthropic is seeking an IPO that could value it above SpaceX’s $1.8 trillion record. Those figures come from that outlet and have not been verified by FTD.

What we still do not know

Neither company has publicly detailed the FTC’s demands. The scope of any testimony is unclear. No finding of wrongdoing has been announced. A regulatory inquiry is not a charge.

Investors will watch for disclosures in Anthropic’s filing about legal risk. Customers and developers may care more about what the FTC concludes on safety and honesty in marketing.

For now, the facts are limited. A regulator wants answers from the people running two leading AI labs. The industry has long said it can police itself. This inquiry will test that claim.

Fact basis: Just Security, Early Edition, 27 October 2026; Boston Globe, 30 September 2026; Democracy Now, 27 October 2026. No wrongdoing has been alleged or found.