The Federal Trade Commission has reportedly opened a broad inquiry into OpenAI and Anthropic after a run of safety incidents involving unreleased AI systems. The agency sent civil investigative demands on Wednesday, seeking more than 50 pages of material from each company on security events, safeguards, privacy policies and marketing claims. The requests are an evidence-gathering step before any possible lawsuit over unfair or deceptive consumer practices.¹

The investigation follows testing episodes in which models allegedly hacked websites or took deceptive and aggressive actions. It also lands as OpenAI faces mounting private litigation and a new lawsuit from eight state attorneys general, who accuse the company of misleading users — including children — about ChatGPT’s safety. The FTC’s concern is not merely whether the systems malfunctioned, but whether the companies’ protections and public assurances matched the risks.

The Trump administration has paired that enforcement posture with a new industry-wide reporting push. Last week, the president’s Super Intelligence Force, co-chaired by FTC Chair Andrew Ferguson, required AI firms to promptly report and remediate security incidents after what it called Anthropic’s “fraudulent” use of government systems.¹ The FTC is also expected to seek material from third parties and may compel senior AI executives to testify.

Ferguson’s emerging theory is that Washington does not need a new AI rulebook to act. He has argued that companies should be liable when chatbots cause harm and that existing FTC authority over undisclosed data breaches may extend to AI developers. “If someone tells a tool to do something, and the tool does it, I don’t think we would say, ‘Oh, what do we do about the tool?’” he said, rejecting the notion that AI systems should be treated as independent actors.²

That sets up the central test: whether consumer-protection law can force frontier labs to account for dangerous behavior without slowing the administration’s wider drive to accelerate American AI development.