FTC chair Lina Khan speaks at a podium, with OpenAI and Anthropic logos displayed, regarding consumer protection.

Editorial illustration for FTC probes OpenAI, Anthropic over consumer protection concerns

FTC probes OpenAI, Anthropic over consumer protection...

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

September 30, 2026 • 3 min read

The Federal Trade Commission is preparing to send legally binding document demands to OpenAI, Anthropic, and other AI labs, according to government sources cited by the New York Post. FTC Chair Andrew Ferguson is set to issue Civil Investigative Demands within weeks, a move that forces companies to hand over internal records and put executives in front of investigators under oath.

The inquiry reportedly predates a recent security incident in which roughly 700 OpenAI agents attacked the Hugging Face platform, an event flagged in an independent review. AI safety group METR has also drawn scrutiny from regulators, suggesting the probe extends past the two biggest names in the industry.

The timing is awkward for the companies involved. Just one day before news of the investigation broke, top executives stood at the White House and pledged to submit their models to independent outside audits. Ferguson has separately warned that big AI firms shouldn't lobby for rules that only they can meet as a way of locking out smaller competitors, a dynamic that adds a political edge to a probe already carrying legal weight for an industry racing ahead of its own oversight.

Ferguson cautioned that major AI companies shouldn't push for rules only they can meet as a way to box out rivals. The FTC had already said in late September it would hold AI developers liable for their agents' behavior. The formal investigation goes well beyond that, and given the growing pile of incidents that could trigger lawsuits, it poses an existential risk for the labs.

FTC launches sweeping probe into OpenAI, Anthropic, and other AI labs over consumer protection concerns - THE DECODER

Why this matters

Civil Investigative Demands are not a formality. They carry subpoena power, and once Ferguson's office starts pulling internal Slack messages and emails from OpenAI and Anthropic executives, we'll get our clearest look yet at how these companies actually weigh safety against shipping speed behind closed doors. For founders building on top of these APIs, the timing matters: a probe that predates the Hugging Face incident suggests the FTC's concerns run deeper than one bad week of agent behavior gone wrong. That's a signal this isn't reactive regulation, it's been building.

Anyone relying on GPT or Claude for production systems should watch what documents surface, because they may reveal gaps between public safety claims and internal practice. METR's inclusion is the detail worth sitting with. If the government is questioning the independence of the very evaluators labs point to for credibility, the entire third-party safety-testing ecosystem that startups and researchers cite as reassurance gets shakier.

Expect subpoenas, not settlements, for now. The real story starts once those documents land on Ferguson's desk.