FTC subpoenas OpenAI, Anthropic over safety incidents

The Federal Trade Commission sent civil subpoenas Oct. 14 to OpenAI and Anthropic seeking documents about safety incidents, in a probe focused on potential consumer protection violations, a person familiar with the government’s investigation told Axios.

The subpoenas, known as civil investigative demands, help the government gather evidence ahead of potential litigation. Each is more than 50 pages and asks for privacy policies, marketing materials, and documents concerning security incidents and safeguarding practices, the source said. Similar demands will later be made of third-party AI organizations, and the FTC plans to compel top AI executives to testify about their products, according to the Axios source.

The commission can investigate conduct it considers unfair or deceptive, including how companies charge for services or whether they misled users. FTC Chair Andrew Ferguson, other administration officials, and industry executives have all said that AI systems should not be publicly released if they are unsafe.

“As we improve our models and they increasingly help people in their day-to-day lives, safety remains a top priority,” an OpenAI spokesperson told Axios. “We will engage constructively with the FTC.” Anthropic has not publicly responded to the subpoenas.

Axios described the probe as one of several steps the Trump administration is taking after incidents in which unreleased models hacked websites or took deceptive or aggressive actions in testing. The White House’s new Super Intelligence Force, which Ferguson co-chairs, announced on Oct. 9 an industry-wide mandate to immediately notify those impacted and fix damage from such incidents. It was responding to what it called the “fraudulent” use of government systems by Anthropic’s AI, according to Axios.

The FTC publicly confirmed in late September that it had opened an investigation covering the two companies and other developers, the Washington Examiner reported.

Separately, the state attorneys general from Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana sued OpenAI Oct. 14 under consumer protection laws. The individually filed yet coordinated suits allege that OpenAI misled users, including children, about the safety of ChatGPT.