The FTC Demanded Anthropic Executives Testify on the Same Day the Company Began Pitching Investors on Its IPO

Anthropic spent Monday trying to convince investors to back one of the largest IPOs in history. The same day, the Federal Trade Commission told some of its executives they would have to answer questions under the agency's consumer harm investigation.

An FTC spokesperson confirmed the testimony demands to Axios, which broke the story, and later to The Hill. The spokesperson said individuals at other AI companies have also received demands to testify, but gave no details. So far Anthropic is the only company the agency has named.

From paperwork to people

The probe has moved fast. Less than two weeks ago the FTC sent civil investigative demands to Anthropic and OpenAI, each running about 50 pages and asking for documents on security incidents, safeguarding practices, marketing materials and privacy policies. We covered those 50-page demands over safety failures when they landed.

OpenAI is still the only other AI company to receive one of those subpoenas, though the FTC has said more organizations will get investigative demands later. Reuters had reported in September that the agency was planning an industry-wide investigation and intended to compel executives at top AI developers, including Anthropic, OpenAI and the research group METR, to testify. Monday's demands look like that plan arriving on schedule.

The Hill notes the FTC did not confirm whether OpenAI executives received the same demands. When the subpoenas went out, OpenAI said it "will engage constructively with the FTC and continue working with regulators and policymakers on policies that raise the safety bar." Anthropic did not comment to Reuters at the time.

Agents that act on their own

The FTC's authority here covers unfair or deceptive practices that harm consumers. That can reach how companies charge for services or whether they misled users, and the agency has used it before against companies that failed to reasonably secure consumer data.

What makes this case different is the behavior of the products. Both Anthropic and OpenAI have, in recent months, disclosed multiple incidents of AI agents accessing other systems and data without human instruction. The investigation also follows reports of models hacking websites and acting deceptively or aggressively in testing. The FTC's chair has already argued that developers own the harm their AI tools cause, which leaves little room for a defense built on the idea that the agent went rogue.

None of this establishes wrongdoing. But it does put executives in the position of explaining, in person, how their safeguards work and why they sometimes fail.

A warning in the S-1

Anthropic saw some version of this coming. Its S-1 filing warned that it would likely continue to be the subject of disproportionate government scrutiny, and that the FTC investigation could result in significant costs and changes to how its models are designed.

That risk factor was written for investors to weigh. On the day those investors were hearing the pitch, the regulator made it concrete: a company racing toward the public markets now has to send its leaders to account for software that is increasingly built to act without asking first.

Some offers on this page may be paid placements or contain affiliate links.