1. United States
  2. Calif.
  3. Letter

Oppose Federal Preemption of State AI Laws

To: Sen. Padilla, Sen. Schiff, Rep. Chu

From: A constituent in Pasadena, CA

December 26

I urge you to oppose any executive order or legislation that would preempt state AI regulation and replace it with centralized federal control. While state-by-state approaches may create complexity, federal preemption threatens innovation and represents a fundamental misunderstanding of how rapidly evolving technology should be governed. States like California, Colorado, Texas, Florida, Vermont, and New York are currently experimenting with different approaches to AI safety, transparency, data privacy, and algorithmic accountability. This represents federalism working as designed, with states functioning as laboratories of democracy. These varied approaches allow us to learn what works before imposing a single national standard that could freeze innovation for decades. History demonstrates the dangers of premature federal regulation. Federal licensing of broadcast frequencies created a rigid media environment dominated by three networks until cable television emerged outside FCC jurisdiction. Nuclear power regulation froze that industry for decades, potentially preventing climate solutions. Every major technology from electricity to the internet faced prophets of doom, yet thrived when allowed to develop without centralized control. The concept of regime uncertainty, developed by economist Robert Higgs, explains how unpredictable government intervention discourages private investment. Businesses already face the EU's AI Act, China's Generative AI Measures, and conflicting state laws. Adding federal centralization would compound this uncertainty rather than resolve it. AI evolves weekly or daily, shaped by billions of user queries and millions of corrections. No central authority can match this distributed intelligence. When AI products fail, markets react immediately through user abandonment and competitive pressure. This creates regulation by consent that's agile enough to match AI's speed. Where AI causes real harms like fraud, impersonation, or privacy breaches, existing law already provides remedies. Civil courts can adjudicate responsibility based on real cases rather than speculative fears. I ask that you reject federal preemption and allow states to continue their experimentation, protecting innovation while addressing genuine harms through existing legal frameworks.

Share on BlueskyShare on TwitterShare on FacebookShare on LinkedInShare on WhatsAppShare on TumblrEmail with GmailEmail

Write to Alejandro Padilla or any of your elected officials

Send your own letter

Resistbot is a chatbot that delivers your texts to your elected officials by email, fax, or postal mail. Tap above to give it a try or learn more here!