Congress just dropped its most ambitious AI legislation yet. The Great American AI Act — a 269-page bipartisan discussion draft released June 4, 2026 — would establish the first comprehensive federal framework for frontier AI development, preempt state AI laws for three years, and create criminal penalties for AI-assisted identity fraud.
The draft was released by Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) alongside four co-sponsors spanning both parties. It’s a discussion draft — not yet a bill with a vote scheduled — but its scope and bipartisan backing make it the most credible federal AI governance attempt in the 119th Congress.
The Three-Year State Preemption That’s Already Causing a Fight
The most controversial section is Section 4: a three-year moratorium on state laws governing AI development. States would retain authority over how AI systems are used within their borders, but lose the ability to set rules on how those systems are built.
The timing is brutal for Colorado. The Colorado AI Act — the most comprehensive state AI law in the country — takes effect June 30, 2026, just 25 days away. If the Great American AI Act passes, Colorado’s development-side provisions could be immediately preempted.
Colorado’s law requires developers to conduct impact assessments, disclose AI use in high-stakes decisions, and give consumers the right to appeal AI-driven outcomes. Big Tech companies have quietly lobbied for federal preemption precisely to avoid a patchwork of similar laws in all 50 states.
What the Bill Actually Requires
The 269-page draft organises around four pillars: frontier AI governance, workforce impact monitoring, cybersecurity standards, and AI research and development investment.
On governance, any AI system meeting a $500 million revenue or compute threshold would need to submit to a new federal review process — similar in concept to the EU AI Act’s high-risk category, but administered by a proposed Center for AI Standards and Innovation funded at $100 million per fiscal year.
The criminal provision is new territory for US AI law. The bill would create penalties for using AI to impersonate individuals without consent — targeting deepfake-based fraud, not artistic parody. It specifically references “AI-facilitated identity fraud” as a federal offense.
Who’s Pushing Back — and Why
The AFL-CIO issued a hard rejection within hours of the draft’s release. The union argues the bill focuses almost entirely on industry governance while providing minimal protections for workers displaced by AI automation. The workforce monitoring section requires data collection on AI’s impact on employment — but critics say monitoring without binding protections is insufficient.
Public Citizen called the state preemption “a gift to Silicon Valley,” arguing it strips state consumer protection tools before any equivalent federal replacement is in place. The Electronic Frontier Foundation flagged the bill’s facial recognition and biometric provisions as potentially too broad.
On the other side, the Computer & Communications Industry Association and several major AI companies have signalled support for federal preemption — citing regulatory certainty as a precondition for continued US AI investment.
Where This Goes Next
Discussion drafts don’t automatically become legislation. This one needs committee hearings, markup sessions, and a floor vote — none of which are scheduled yet. But the bipartisan co-sponsorship (six representatives from four states) is a stronger foundation than most AI bills have reached.
The most likely near-term outcome is a summer hearing in the House Science, Space, and Technology Committee, where Obernolte serves as chair. A floor vote before the August recess would be aggressive; a 2027 timeline is more realistic given the legislative calendar.
Frequently Asked Questions
What is the Great American AI Act?
The Great American AI Act is a 269-page bipartisan discussion draft federal AI bill released June 4, 2026, by Reps. Obernolte and Trahan. It proposes the first comprehensive federal framework for frontier AI governance, including a three-year preemption of state AI development laws and a new Center for AI Standards and Innovation.
What does “state preemption” mean in the Great American AI Act?
State preemption means the federal law would override conflicting state laws. Under this bill, states would keep authority over how AI is used within their borders, but lose the power to regulate how AI systems are developed. This directly affects laws like the Colorado AI Act (effective June 30, 2026).
Which AI companies would be affected by the bill?
Any AI company or product meeting a $500 million revenue or compute threshold would face federal review requirements. This primarily targets frontier model developers — OpenAI, Anthropic, Google DeepMind, Meta AI, and Microsoft — while likely exempting smaller AI startups below the threshold.
Does the Great American AI Act protect workers from AI displacement?
The bill requires federal data collection on AI’s impact on the US workforce, but the AFL-CIO and other labor groups argue this falls far short of actual worker protections. There are no binding requirements on employers to notify, retrain, or compensate workers displaced by AI in the current draft.
When could the Great American AI Act become law?
It is currently a discussion draft with no scheduled committee vote. Realistically, a floor vote could come in late 2026 or 2027 after committee hearings and markup. The bill’s bipartisan backing gives it a stronger path than most AI legislation, but the state preemption debate will likely require significant revision before it advances.
This bill arrives days after President Trump’s AI executive order, which took a lighter regulatory approach — the Great American AI Act represents Congress asserting its own lane. Our coverage of Trump’s AI executive order covers the current White House position in detail. For the broader context of how AI is reshaping work, see our analysis of AI replacing jobs in 2026. On the OpenAI side, Congress is also watching the OpenAI Frontier Governance Framework closely as an industry self-regulation benchmark.
Last Updated: June 2026
