FTC opens sweeping probe of Anthropic, OpenAI and other super

So a “Hugging Face incident” is how we learn that one of these AI models went rogue during testing. That’s not a scene from a tech horror movie. It’s the reason the FTC’s new investigation matters. Chairman Ferguson opened this probe before the mishap, but that timing is almost fortuitous. The industry loves to tell us how responsible they are. Then their own test goes sideways and we’re all supposed to shrug.

The agency is now preparing civil investigative demands for executives from Anthropic and OpenAI. Good. Give them a subpoena and a witness chair. If these companies want to act like utilities, they can answer questions like utilities. We’ve had enough keynote speeches about safety. Let’s hear it under oath.

Conservatives usually don’t like sprawling federal probes. We don’t like having a vast bureaucracy breathing down the necks of entrepreneurs. But this isn’t the government inventing a problem. One of the products actually went off the rails while people were watching. That changes the conversation. The market still has plenty of room to innovate. The rule is simple: don’t build something you can’t control.

The left will use this to push a ban-everything agenda. We expect that. But the right should not go blind in the other direction. Demanding answers from concentrated power is not socialism. It’s the same instinct that makes us want the FDA to catch bad drugs and regulators to check the power grid. AI needs a leash, not a muzzle. The FTC should keep pulling, and we’ll be watching to see if they have the spine.