U.S. lawmakers release Great American AI Act draft as AI regulation fight intensifies

U.S. lawmakers release Great American AI Act draft as the fight over AI rules intensifies

Image credit: “Technology” image from the official press release page of Rep. Jay Obernolte’s office. Image format: JPG, not SVG. Source: https://obernolte.house.gov/media/press-releases/obernolte-trahan-release-discussion-draft-great-american-ai-act

WASHINGTON, June 5, 2026 — A new artificial intelligence bill draft in the United States has become one of the most closely watched AI policy stories of the day because it targets a central question in AI governance: should AI rules be set mainly by the federal government, or should individual states keep broad authority to regulate the technology on their own?

According to Reuters, Republican Rep. Jay Obernolte and Democratic Rep. Lori Trahan released a bipartisan discussion draft titled the Great American Artificial Intelligence Act of 2026. The proposal seeks to create a federal framework for AI governance, with a particular focus on advanced frontier AI models. It would also limit state and local authority to regulate the development of AI models directly.

What happened?

On June 4, 2026, Rep. Jay Obernolte’s office published the discussion draft of the Great American AI Act. The text is not yet enacted law. It is a discussion draft intended to gather feedback from stakeholders, experts and the public before formal introduction in the House.

The official announcement describes the draft as a bipartisan effort to govern AI effectively, protect Americans from emerging risks and preserve U.S. leadership in AI innovation. Obernolte, Trahan and several other lawmakers were listed as supporters of the public-feedback effort.

The most contentious provision concerns federal preemption. The draft would block states and local governments from enforcing or creating laws that specifically target AI model development. At the same time, the section-by-section summary says the draft does not preempt generally applicable laws, common law remedies, or laws regulating AI use or deployment.

What would the bill do?

The draft’s first major pillar is frontier artificial intelligence governance. It defines key terms including AI model, foundation model, frontier model, frontier developer, large frontier developer, catastrophic risk, critical safety incident and independent verification organization.

The proposal would formally establish the Center for AI Standards and Innovation within the U.S. Department of Commerce. According to the official summary, the center would develop voluntary AI security guidelines, best practices and standards covering areas such as adversarial robustness, interpretability, supply-chain threats, model tampering, synthetic-content detection and independent verification licensing.

Large frontier AI developers would have to publish frontier AI frameworks, report details about new model deployments, and notify authorities about critical safety incidents. The official summary says certain violations could lead to fines of up to $1 million per day.

Why is this news important?

The draft arrives while the United States still lacks a comprehensive federal AI law. In the meantime, many states have been moving ahead with their own AI-related rules. Technology companies often favor a national standard, arguing that a patchwork of state laws creates heavy compliance burdens. Civil society groups, however, argue that states have been the fastest source of protection against AI harms.

Reuters reported that the draft was welcomed by parts of the technology industry because it could create a unified national framework. But the same report also noted criticism from consumer advocates, who warned that it could weaken state-level oversight of AI.

Why are advocacy groups pushing back?

Public Citizen sharply criticized the Obernolte-Trahan draft. The organization said the proposal would prevent state and local governments from regulating AI model development while leaving oversight largely to a federal system that has struggled for years to pass strong AI protections.

Public Citizen also argued that the draft does not adequately address algorithmic discrimination, housing discrimination, employment discrimination, consumer fraud, youth mental-health harms, AI companions, deepfake exploitation and growing market concentration among large technology companies.

The ACLU also criticized the proposal. It warned that, if enacted, the bill could block states from enforcing or creating protections related to privacy, anti-discrimination, AI safety and other areas where AI already affects employment, healthcare, lending, education and civil rights.

What could change if the draft advances?

If the proposal moves forward in Congress, it could become one of the most significant U.S. attempts to put frontier AI under a federal legal framework. The bill is not limited to model safety. It also includes provisions on AI literacy, education, labor-market research, worker displacement, cybersecurity, critical open-source software, public data, AI research resources, international standards and data-center cooling.

For AI companies, the most important part may be transparency and independent verification for frontier models. For workers, the most relevant provisions include labor-market forecasting, AI-sensitive occupations, and new reporting language when AI is a substantial factor in a qualifying mass layoff.

The hardest question remains political as much as technical: how to preserve AI innovation while building credible safeguards for real-world harms from increasingly powerful AI systems.

What comes next?

The next issue to watch is whether the draft is formally introduced as a bill in the House, and whether the state-preemption language changes after feedback from civil-liberties groups, consumer advocates, state officials and industry stakeholders.

Because this is still a discussion draft, the language may change substantially. Even so, the news marks a new phase in the U.S. AI debate. The argument is no longer only about which AI model is more capable. It is about who gets to write the rules for AI: Congress, the states, regulators, or the technology companies building the systems.

Verified sources

1. Reuters — “US House lawmakers release draft bill to prohibit state AI rules”: https://www.reuters.com/business/us-house-lawmakers-release-draft-bill-regulate-ai-2026-06-04/
2. Office of Rep. Jay Obernolte — Official press release on the Great American AI Act discussion draft: https://obernolte.house.gov/media/press-releases/obernolte-trahan-release-discussion-draft-great-american-ai-act
3. Great American Artificial Intelligence Act of 2026 discussion draft: https://obernolte.house.gov/sites/evo-subsites/obernolte.house.gov/files/evo-media-document/the-great-american-ai-act-discussion-draft-website-compressed-compressed.pdf
4. Section-by-section summary of the discussion draft: https://obernolte.house.gov/sites/evo-subsites/obernolte.house.gov/files/evo-media-document/gaaia-discussion-draft-section-by-section-website.pdf
5. Public Citizen — Response to the Obernolte-Trahan draft: https://www.citizen.org/news/obernolte-trahan-bill-strips-states-authority-to-protect-consumers-workers-and-children/
6. ACLU — Response to the draft AI bill that would preempt state laws: https://www.aclu.org/press-releases/aclu-reacts-to-draft-bipartisan-ai-bill-that-would-preempt-state-laws