1. United States
2. Calif.
3. Letter

I urge you to oppose the provision in the current budget reconciliation bill that states: “No State or political subdivision thereof may enforce any law or regulation regulating artificial intelligence models, artificial intelligence systems, or automated decision systems during the 10-year period beginning on the date of the enactment of this Act.” This clause is an unconstitutional overreach. It strips states of the power to protect their residents from emerging harms caused by AI systems and offers no federal regulatory framework in return. 1 The Tenth Amendment reserves powers not granted to the federal government to the states. Nowhere does the Constitution give Congress exclusive authority over AI regulation. New York v. United States (1992) makes clear that the federal government cannot prohibit states from acting in areas of retained sovereignty. This provision violates that principle. 2 Most AI use cases—such as in hiring, housing, education, healthcare, and policing—have local impacts and fall under state authority. In United States v. Lopez (1995), the Supreme Court ruled that Congress cannot use the Commerce Clause to regulate issues where the link to interstate commerce is too weak. This provision uses similar overreach to block state action. 3 The anti-commandeering doctrine prohibits the federal government from forcing or forbidding states to legislate on particular issues. Murphy v. NCAA (2018) and Printz v. United States (1997) confirm that Congress cannot bar states from enacting or enforcing their own laws. This clause directly violates that doctrine. 4 Under the Supremacy Clause, federal law may only preempt state law when Congress establishes a comprehensive federal framework or demonstrates a compelling need to override state action. This provision does neither. Arizona v. United States (2012) upheld federal preemption only when a full federal system existed—in that case, immigration. There is no equivalent AI system here. 5 Allowing companies to develop and deploy AI systems without oversight for a decade is dangerous. AI is already being used in: ◦ Predictive policing and surveillance ◦ Hiring and housing decisions ◦ Insurance scoring and medical diagnosis ◦ Student monitoring and autonomous vehicles These systems have shown racial, gender, and economic bias, often operate without transparency, and can produce unaccountable outcomes. States must retain the power to intervene to protect their residents. 6 AI will transform nearly every sector of life in the coming decade. Regulation must evolve in parallel—not be suspended for ten years. State and local governments are best equipped to respond to community needs, enforce civil rights, and ensure accountability in new technologies. Preventing them from acting will cause irreparable harm. This provision is unconstitutional, unjustified, and dangerous. I respectfully urge you to oppose it. Our communities deserve strong protections.