A new blueprint lays out how the U.S. can build a lasting federal framework for governing powerful AI systems. The plan has three parts: create a national framework that builds on emerging state laws like California’s SB 53 and New York’s RAISE Act; strengthen the newly formed CAISI as the federal government’s main safety institution; and launch a broader resilience plan to tackle national security and public safety threats from advanced AI. The White House’s recent executive order on AI innovation and security adds momentum. The key idea is to weave together state efforts and federal action into a system that can adapt as tech evolves—without stifling progress. Why it matters: This is a rare opening to embed democratic guardrails into AI before it outpaces policy, turning a patchwork of state rules into a coordinated national defense. The goal: smart, flexible oversight that evolves with the tech itself.