The Federal Government Moves to Preempt State AI Laws as Congress Drafts National Framework

The United States is rapidly approaching a constitutional and regulatory inflection point over who governs artificial intelligence. After years of congressional gridlock allowed individual states to pass their own AI statutes, the federal government is now executing a coordinated offensive to seize control of the regulatory landscape.[2]

The primary mechanism is federal preemption—a legal doctrine that would invalidate state-level AI laws in favor of a unified national standard. This push is advancing simultaneously through executive orders, Department of Justice litigation threats, and a sweeping new bipartisan bill introduced in the House of Representatives.[1][3][5]

For enterprise compliance teams and frontier AI developers, the stakes are existential. Companies are currently navigating a fragmented patchwork of state regulations that took effect in early 2026, including California's transparency mandates and Texas's restrictions on government AI use. The federal government argues this state-by-state approach stifles innovation and creates impossible compliance burdens for startups.[2][7][8]

The executive branch is actively preparing to challenge state AI laws in federal court. The foundation for this offensive was laid in December 2025, when President Donald Trump signed Executive Order 14365, titled "Ensuring a National Policy Framework for Artificial Intelligence."[8]

The order explicitly criticized "excessive State regulation" and established a policy of maintaining U.S. global AI dominance through a "minimally burdensome" national framework. Crucially, it directed the U.S. Attorney General to establish a specialized legal unit to challenge state laws that conflict with this federal posture.[3][8]

On January 9, 2026, Attorney General Pam Bondi formally launched the DOJ's AI Litigation Task Force. According to internal DOJ memoranda, the Task Force is authorized to challenge state AI statutes on the grounds that they unconstitutionally regulate interstate commerce or are preempted by existing federal regulations.[3]

In March 2026, the White House reinforced this posture by releasing a seven-pillar "National Legislative Policy Framework for Artificial Intelligence." The blueprint explicitly urged Congress to pass a unified federal rulebook that includes targeted preemption of state AI laws deemed unduly burdensome.[2]

Meanwhile, Congress is coalescing around a comprehensive federal standard targeting "frontier" models. On June 4, 2026, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a 269-page bipartisan discussion draft of the "Great American Artificial Intelligence Act of 2026" (GAAIA).[1][5]

If enacted, the GAAIA would create the first comprehensive federal AI governance regime in the United States. The legislation primarily targets "large frontier developers"—defined as companies with more than $500 million in annual revenue that train advanced AI models, a threshold that captures industry leaders like OpenAI, Anthropic, Google, and Meta.[1][5]

The draft bill's most consequential provision is a three-year preemption clause that would freeze all state laws specifically regulating the development of AI models. However, it explicitly preserves state laws of general applicability, meaning existing consumer privacy laws like the California Consumer Privacy Act (CCPA) would remain in effect.[5][7]

The GAAIA also introduces stringent transparency and auditing requirements. Frontier model developers would be required to disclose model information, obtain third-party audits through designated Independent Verification Organizations, and report critical safety incidents to federal authorities. Violations could trigger fines of up to $1 million per day.[5]

Beyond model governance, the bill includes robust whistleblower protections for employees who report AI violations at any company, not just frontier developers. It also establishes an AI Workforce Research Hub within the Department of Labor to track the technology's impact on the labor market.[4]

The chilling effect of the DOJ Task Force and pending federal legislation is already reshaping state-level policymaking, with some states retreating or recalibrating in the face of federal pressure.[7]

The most prominent example occurred in Colorado. The state had originally passed an ambitious AI Act (SB 24-205) that imposed a strict risk-based framework on developers and deployers, requiring annual impact assessments and algorithmic discrimination prevention. However, after facing intense industry pushback and the looming threat of federal preemption, Governor Jared Polis signed SB 26-189 on May 14, 2026, repealing the original law entirely.[2][7]

The replacement law strips out the mandatory risk management programs and impact assessments, replacing them with lighter obligations centered on pre-use consumer notices and 30-day adverse-outcome explanations.[2][7]

Meanwhile, California's Transparency in Frontier AI Act (SB 53) and the Texas Responsible AI Governance Act (TRAIGA) both took effect on January 1, 2026. California's law requires developers of large models to publish risk frameworks and implement safety reporting, while Texas's law imposes categorical bans on AI systems designed for behavioral manipulation or unlawful discrimination, though it was narrowed to primarily cover government use.[2]

Despite the aggressive posturing, the path forward remains highly contested. The DOJ's AI Litigation Task Force has not yet filed any formal challenges against state laws, leading some legal analysts to question whether the executive branch is waiting for Congress to act first.[6]

The GAAIA also faces significant hurdles. While it enjoys bipartisan backing, the preemption clause has drawn fierce opposition from state-rights advocates and civil society groups who argue that states must retain the authority to protect their citizens from algorithmic harm. With Congress scheduled to break for recess in August 2026, the window to pass a unified federal framework is rapidly closing.[7]