The White House released its first national AI policy framework on March 20, with officials urging Congress to turn it into law "this year."
Michael Kratsios, director of the White House Office of Science and Technology Policy, and David Sacks, the administration's AI and crypto czar, presented the framework as a legislative blueprint. The central argument: a single federal standard should replace the growing patchwork of state-level AI laws.
Federal vs. State Control
The framework draws a clear line. States would lose authority to regulate AI development itself, which the administration treats as an interstate and national security matter. But states keep their power over consumer protection, child safety, fraud prevention, and zoning. The framework also says states shouldn't penalize AI developers for how third parties misuse their tools.
"We need one national policy, not a 50-state patchwork of laws," Kratsios said.
Child Safety and Copyright
On child safety, the framework calls for age-verification mechanisms, parental controls for privacy and screen time, and features that reduce risks of sexual exploitation and self-harm. It references the bipartisan "Take it Down Act" as a model.
Copyright gets a lighter touch. The administration acknowledges disputes over whether AI companies can train models on copyrighted material, but prefers to let courts resolve those questions rather than writing new legislation. The framework says it supports protecting creators' intellectual property while preserving "lawful innovation."
Energy Costs
One concrete requirement stands out: a "Ratepayer Protection Pledge" that would make AI companies fund their own power infrastructure for data centers and absorb the costs, rather than passing electricity price increases to consumers. Companies would also commit to local hiring and job training.
The framework is a proposal, not a law. Congress would need to act on it, and the "this year" timeline is ambitious given how slowly AI legislation has moved so far. But the preemption push matters right now. States like California, Colorado, and Illinois have been advancing their own AI bills. If a federal standard passes, much of that work could be overridden.