Krishnan Tells FT No 'FDA for AI' Under Trump, Blames Doomers

The person who has been shaping the White House's AI posture since January 2025 just used his exit interview to close a door. Sriram Krishnan, the senior AI policy adviser who departed at the end of June, told the Financial Times that "there will not be an FDA for AI" under Trump, and blamed the growing AI backlash on the industry's own "doomer" messaging.

That is a definitive line. For months the possibility of a horizontal frontier-model regulator, a pre-approval body that would treat large models a bit like drugs, has been an implicit hedge in Washington conversations. Krishnan is saying, plainly, the president does not want it. His broader framing, that "This notion of imminent AGI has been a distraction and harmful and now effectively proven wrong," dismisses the safety-lobby argument he thinks put the idea on the table in the first place. He has previously characterised the "EA/AI safety/'doomer' lobby" as now being "out of power."

For practitioners the practical read is straightforward. The federal regulatory ceiling on frontier labs under this administration is the voluntary framework Trump signed on June 2, 2026, which asks AI companies to submit their most powerful models for government testing up to 30 days before public release. That is it. There is no binding pre-approval, no mandatory audit, no single agency of jurisdiction, and Krishnan is telling you not to expect one. Product roadmaps and legal reviews should be calibrated to that, not to a hypothetical FDA-style gate. The "FDA for AI" phrase itself had briefly been floated inside the administration by NEC director Kevin Hassett, only for a draft framework to be pulled an hour before signing.

The honest caveat is that this is one departing adviser's view, delivered on the way out the door, and administrations shift after high-profile incidents. What the reporting does not give you is what happens if a serious model-caused harm actually occurs on this watch, or how much of "no FDA for AI" applies to sector-specific rules in health or aviation versus a horizontal regulator. It also leaves open who succeeds Krishnan and whether the outside institution he plans to start becomes the real centre of gravity for pro-industry AI policy in Washington.

For US frontier labs and their investors, this is about as clear a signal as they were going to get that binding federal rules are not coming this administration. The interesting political fight now moves to the states, and to whoever fills the vacuum Krishnan is walking into on his way out.