On March 20, the White House published its long-awaited National AI Legislative Framework. A four-page blueprint urging Congress to adopt a federally unified, innovation-oriented AI regime. The centerpiece: preempting state AI laws in favor of a "light-touch" federal approach.

The problem? Congress is deadlocked, and states aren't waiting around.

What the Framework Actually Says

The framework targets for preemption any state law that contradicts the administration's strategy for "global AI dominance," regulates AI development, restricts Americans' ability to use AI for activities that would be legal without AI, or penalizes AI developers for third-party misuse of their models.

Translation: the White House wants to make it very hard for states to regulate AI companies. The exceptions are narrow. child safety, fraud prevention, consumer protection, and state procurement rules can stay.

The State-Level Revolt

As NPR reported, states have already passed dozens of AI laws covering everything from child safety to deepfake disclosure to hiring algorithm transparency. They're not doing this because they love regulation. They're doing it because the federal government has left a vacuum.

And now the administration wants to override all of that with a framework that explicitly prioritizes "global AI dominance" over comprehensive safety standards. The tension is real and unresolved.

Our Take

A unified federal AI framework makes sense in theory. The patchwork of state laws creates real compliance headaches for companies operating nationally. But a four-page document that essentially says "don't regulate us" isn't a framework. It's a wish list. And as long as Congress can't agree on anything, states will keep filling the gap with their own rules. Expect the legal battles over preemption to last years.

Framework reported by POLITICO. State response coverage by NPR.