Who writes America's AI rules is still being decided

Not a comprehensive one. Federal AI policy runs mainly through executive orders, agency guidance and voluntary commitments, such as the June 2026 order on pre-release testing of frontier models and the September 2026 White House Accord on Super Intelligence. Narrower federal laws, like the TAKE IT DOWN Act on nonconsensual intimate images, and existing consumer protection, civil rights and sector laws also apply to AI.

Yes. An executive order cannot strike down a state statute; only Congress or the courts can preempt state law, and as of November 2026 neither has displaced a state AI law. The Justice Department joined a challenge to Colorado's original AI Act in 2026, and Colorado replaced that law with a narrower one that takes effect Jan. 1, 2027.

Democrats take control of both chambers in January 2027. Reported priorities include protecting consumers from data center electricity costs, independent testing of advanced AI systems and clearer liability for AI harms, along with hearings and oversight. Any bill needs the president's signature or a two-thirds vote in both chambers to override a veto, so oversight is likely to move faster than new law.

It depends on where you operate and how you use AI. The laws drawing the most compliance work include California's frontier AI transparency law (SB 53), the Texas Responsible Artificial Intelligence Governance Act, Colorado's automated decision-making law from Jan. 1, 2027, and New York's RAISE Act for the largest frontier developers. Chatbot, deepfake, health insurance and education laws are spreading fastest across states.