FTC subpoenas Anthropic and OpenAI over artificial intelligence safety concerns

Want to bookmark your favourite articles and stories to read or reference later? Start your Independent Membership today.

The U.S. Federal Trade Commission has issued civil subpoenas to AI developers Anthropic and OpenAI seeking records on safety incidents, Axios reported, citing a person familiar with the matter.

The probe comes amid mounting concerns over AI safety after models reportedly hacked websites and acted deceptively or aggressively in tests, underscoring the dangers of increasingly autonomous systems and the adequacy of existing protections.

Reuters reported last month that the agency was preparing an industry-wide inquiry.

The FTC also intends to compel executives at Anthropic, OpenAI, and research group METR to testify.

An OpenAI spokesperson stated the firm "will engage constructively with the FTC and continue working with regulators and policymakers on policies that raise the safety bar."

Neither the FTC nor Anthropic immediately replied to requests for comment from Reuters regarding the Axios report.

The agency’s demands stretch to more than 50 pages each, requesting information on security incidents, protective measures, privacy policies, and marketing materials, according to the news website.

The FTC possesses broad authority to litigate against deceptive or unfair practices, a power it has previously deployed against firms failing to take reasonable steps to secure consumer data.

FTC Chairman Andrew Ferguson told Reuters last month that developers instructing agents in cybersecurity tests leading to hacks should face liability for any harm caused.