Congress wants to freeze every state AI law for three years
A 269-page discussion draft dropped late Thursday that could override every state-level AI regulation in America. Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) unveiled the Great American Artificial Intelligence Act on June 4, and the timing is not subtle. Colorado’s Consumer Protections for Artificial Intelligence Act, the first comprehensive state AI law in the US, takes effect June 30. This bill would freeze it before it starts.
The core mechanism is a three-year preemption of state AI laws related to the development of frontier AI models. California’s pending AI bills, Colorado’s anti-discrimination requirements, Texas and Illinois laws on biometrics and automated hiring. All paused. The bill allows states to regulate AI after deployment under laws of “general applicability,” but anything touching how models get built would be federal territory.
What the bill demands from large AI companies (those with over $500M in annual gross revenue):
- Publish public Frontier AI Frameworks disclosing how they govern their most capable models
- Report critical safety incidents to the federal government
- Allow auditors to verify cybersecurity mitigation plans
- Support a $100M/year Center for AI Standards and Innovation (CAISI) inside the Commerce Department
CAISI is the renamed Biden-era AI Safety Institute. It currently exists via executive order but lacks congressional authorization. The bill would formally codify it at NIST with funding through FY2029. The center would develop voluntary guidelines, evaluate AI systems, and monitor progress. The DOE and NIST would jointly lead international AI standards development, “expressly excluding China.”
Additional provisions include criminal penalties for using AI to impersonate government officials, Census Bureau integration of AI usage questions into federal surveys, extension of the Cybersecurity Information Sharing Act through 2035, GAO evaluation of federal AI adoption, and formal codification of the National AI Research Resource (NAIRR).
The reaction split along predictable lines. Labor unions moved fast. The AFL-CIO, American Federation of Teachers, and Association of Flight Attendants issued a joint statement: “Hard no. This bill is a giveaway to the AI industry.” Over 200 state lawmakers signed a letter opposing preemption. Brad Carson of Americans for Responsible Innovation called it a “generational mistake.”
Tech industry groups ITI and NetChoice praised the bill. Rep. Houchin (R-IN) framed it as preventing America from regulating “ourselves into falling behind China through a patchwork of fifty different state laws.”
The White House has not weighed in. A December 2025 executive order specifically targeted the Colorado AI Act, claiming it would “force AI models to produce false results” through its anti-discrimination provisions. The Great American AI Act reads as the legislative follow-through on that position.
The preemption fight has recent history. A previous 10-year preemption moratorium was stripped by a 99-1 Senate vote. This draft scaled back to three years, but the structural tension is the same: preempting state laws without matching federal protections leaves a gap. Colorado’s law includes actual anti-discrimination requirements for high-risk AI systems in employment, housing, healthcare, and financial services. The federal bill’s “general applicability” carve-out for states is vague enough to litigate for a decade.
This is a discussion draft, not a formally introduced bill. The sponsors are actively soliciting feedback at [email protected] before formal introduction. With Colorado’s law taking effect in 25 days, the legislative clock is real.
Sources: Obernolte press release, FedScoop, AI Tools Recap, BuildFastWithAI