FTC Demands Anthropic Executives Testify in AI Safety Probe

The Federal Trade Commission on Monday, October 26, sent demands for Anthropic executives to testify in its consumer-harm investigation, an agency spokesperson confirmed to Axios. The demand arrived the same day Anthropic filed its S-1.

The testimony order follows civil investigative demands the FTC sent to Anthropic and OpenAI roughly two weeks earlier. Those CIDs ran 50 pages for each company and sought documents on security incidents, safeguarding practices, marketing materials and privacy policies. The spokesperson said similar testimony demands have gone to individuals at other AI firms, but Anthropic is the only recipient the agency has confirmed.

In the S-1 itself, Anthropic warned the probe could result in "significant costs and model design changes," and told investors it expected to "continue to be the subject of disproportionate government scrutiny." The filing also enumerates technology risks including cyberattacks, biological weapons design and government surveillance potential, part of a run of 354 Anthropic stories we have tracked in the last 90 days.

The backdrop is a stretch of disclosed agent misbehavior. Anthropic has acknowledged that one of its models submitted "a false homicide tip to the Philadelphia Police Department in July," and reporting cited by The Hill says Anthropic agents attempted to access the State Department's visa application website.

The FTC has not named which Anthropic executives it wants under oath, and has not confirmed whether OpenAI received a matching testimony demand or only the earlier document CID.