US competition watchdog expands investigation of Anthropic and OpenAI

The US Federal Trade Commission is widening a probe into Anthropic, OpenAI and other AI companies over the harms posed by their products, following a spate of hacking incidents.

The regulator will demand information and seek testimony from top executives, according to a person familiar with the investigation. The FTC is also looking into METR, an AI safety research group that White House advisers have accused of being intertwined with Anthropic’s investors and staff.

The FTC confirmed the probe but declined to comment further. The investigation marks the first official regulatory action focused on rogue AI agents amid fears that the industry could pose outsized risks to humanity.

The widening of the probe, which was first reported by the New York Post, comes a day after President Donald Trump gathered AI bosses to the White House for a high-stakes meeting. Anthropic, OpenAI and four other labs subsequently signed an “accord” to self-regulate their products and to monitor the cyber security and biosecurity threats posed by advanced models.

The Trump administration has insisted that existing laws are sufficient to limit the potential harms caused by frontier models despite a series of alarming incidents in which AI models hacked groups including the Hugging Face platform and the Australian health service.

“A bad actor is going to get caught, and that’s why we have the [Department of Justice], that’s why we have the FBI, that’s why we have the CIA,” Trump said following the White House meeting.

FTC chair Andrew Ferguson, who has previously said the AI companies could be held liable for the actions of their models, was also present at the lunch on Tuesday.

Some lawyers have suggested that the FTC could use violations of the White House accord as evidence to pursue a potential case under the deceptive practices provision in the Federal Trade Commission Act.

The DoJ has signalled some hesitancy to take on AI alongside its competition enforcement counterpart. Attorney-general Todd Blanche told reporters earlier this month that the justice department was made up of “prosecutors, not regulators”.

“So we’re not going to try to regulate AI,” Blanche said.

Trump, however, has said the DoJ and other federal law enforcement agencies will rein in AI “if we have to”.

Anthropic, OpenAI and METR did not immediately respond to requests for comment.

OpenAI is already facing a lawsuit from Florida’s attorney-general, who alleges the company did not do enough to stop its products from harming children. OpenAI said at the time it had “put in place industry leading protections and policies”.

The Florida attorney-general also moved this week to stop OpenAI from further developing its advanced models until third-party safety guardrails are in place.