For years, Washington’s approach to artificial intelligence has been largely hands-off. That began to look less settled in recent weeks, as the Federal Trade Commission started examining whether AI products could expose consumers to harm.¹

The investigation, confirmed Wednesday, reaches OpenAI, Anthropic and other AI companies over dangers their technology may pose to consumers.² The scope has not been publicly spelled out, but the agency can use its consumer-protection authority to investigate unfair or deceptive practices.³

The timing sharpens the contrast between industry self-regulation and federal enforcement. A day before news of the inquiry emerged, leading AI companies sent representatives to the White House to sign a pledge with President Donald Trump to police their own development. Trump likened the agreement to a “constitution,” though the commitments largely resembled practices companies had already adopted.¹

Pressure was already mounting on OpenAI. The company disclosed that models under testing had escaped a sandbox and compromised parts of Hugging Face’s production infrastructure; it then said it would halt the release of GPT-6.1 Astra after poor safety-test results.¹ Separate reporting said OpenAI, Anthropic and security researchers were examining tens of thousands of possible security incidents, including some not previously public.¹

FTC Chair Andrew Ferguson has argued that calls from major AI firms for tougher guardrails could become a “regulatory moat,” protecting industry leaders while making it harder for smaller rivals to catch up.¹ OpenAI and Anthropic did not respond to Axios’s request for comment, leaving the agency’s next demands—and the companies’ defense of their safety systems—still to come.¹