On June 4, 2026, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a discussion draft of the Great American Artificial Intelligence Act (GAAIA), proposing what would become the first comprehensive federal framework for governing AI in the United States.

The bill is bipartisan and ambitious. It covers frontier model governance, workforce impacts, cybersecurity, and international cooperation. It would also preempt state-level AI laws for three years, a provision that is already drawing both praise and sharp opposition before the bill has even been formally introduced.

What the Bill Actually Does

The GAAIA is structured around four main areas:

Frontier AI Governance. The bill requires developers of frontier AI models to disclose information about those models to regulators, obtain independent third-party audits through designated Independent Verification Organizations (IVOs), and refrain from retaliating against employees who raise safety concerns. This is a meaningful compliance burden for the major labs, but it is also an acknowledgment that self-regulation has not been enough.

Workforce. There is a dedicated section on AI’s impact on employment. The bill would direct the Census Bureau and the Bureau of Labor Statistics to update their surveys to track AI use and adoption across the economy, giving policymakers actual data instead of speculation about job displacement.

Cybersecurity. The draft adds federal penalties for using AI to impersonate government officials, an area where voice and image AI have already created real risks.

Research, Development, and International Cooperation. The bill would formally codify the Center for AI Standards and Innovation within the Commerce Department and authorize $100 million per year for fiscal years 2027 through 2029 to support standards work.

The State Preemption Question

The most contested element of the bill is the three-year preemption of state laws that “specifically regulate the development” of AI models. Importantly, the preemption would not apply to laws governing the use or deployment of AI.

This distinction matters. Colorado’s SB 205, which takes effect June 30, 2026, covers high-risk AI deployments. Under the current draft, it would likely remain in force because it targets deployment, not model development. But companies navigating a patchwork of state laws across 12 or more states would get some relief on the development side.

The preemption clause has already triggered pushback. Public Citizen and consumer advocacy groups argue the bill strips states of their ability to protect consumers and workers. The House Democratic Commission on AI released a statement in opposition within hours of the draft’s release. The bill’s path to formal introduction, let alone passage, is uncertain.

This Is Still a Discussion Draft

It is worth being clear about where this actually stands. The GAAIA is a discussion draft, which means Obernolte and Trahan are soliciting feedback from industry, civil society, and other legislators before formally introducing it. There is no co-sponsorship list, no committee markup, and no floor vote scheduled.

Congress has tried to pass comprehensive AI legislation before. It has not succeeded. The GAAIA is more substantive and more bipartisan than most previous attempts, but the political dynamics around state preemption, AI liability, and civil rights protections have killed similar bills before.

That said, the discussion draft signals something real: there is bipartisan appetite in Congress to establish a federal floor before the patchwork of state laws becomes fully operational. Companies that have been waiting for federal clarity before investing in compliance infrastructure may be waiting for another 18 to 24 months at minimum.

What This Means for Business

If you are running a business that uses AI, here is what to watch:

Colorado’s June 30 deadline is still real. Whatever happens with the GAAIA, Colorado’s high-risk AI law takes effect at the end of this month. Companies that deploy AI in consequential decisions affecting Colorado residents need to assess their exposure now, not when federal legislation potentially passes.

The audit requirement is coming for frontier providers. If the bill passes in anything close to its current form, the major AI labs you rely on will face independent third-party audits of their models. That will likely accelerate disclosures about model capabilities, failure modes, and safety practices, giving enterprise buyers better information.

State preemption relief is not guaranteed. Do not build your compliance strategy around the assumption that federal law will wipe out state requirements. Even if the GAAIA passes with preemption provisions intact, state deployment laws like Colorado’s would still apply.

First-mover advantage in AI governance. Companies that build transparent AI governance practices now, regardless of regulatory requirements, are better positioned when auditors, customers, or regulators start asking questions. The GAAIA’s IVO audit framework is a preview of the kind of documentation standards that will eventually become normal.

The Great American AI Act is not law yet, and it may never be. But it is the most serious attempt at federal AI governance the U.S. has produced, and the provisions in the discussion draft tell you where the regulatory conversation is heading.