The White House is calling on Congress to override state-level artificial intelligence laws and adopt a lighter federal regulatory approach, arguing that a unified national framework is needed to prevent a patchwork of rules from slowing innovation and weakening U.S. competitiveness.

The legislative recommendation also wants to preserve limited state authority in areas such as enforcing general consumer protection laws, policing fraud, and regulating zoning for AI infrastructure.

The framework opposes state intervention, saying AI development crosses state lines and is closely tied to national security and foreign policy, making state-by-state regulation inappropriate.

It also argues states should not penalize AI developers for how third parties use their systems or restrict lawful uses of AI technology.

Alongside preemption, the administration is pushing Congress to avoid creating new federal regulatory bodies dedicated to AI.

Instead, it recommends relying on existing agencies with subject-matter expertise and encouraging industry-led standards to guide development and deployment.

The proposal also calls for the creation of regulatory “sandboxes,” controlled environments where companies can test AI applications with fewer regulatory constraints, as a way to accelerate innovation and maintain U.S. leadership in the global AI race.

Officials frame the approach as a way to strike a balance between oversight and growth, emphasizing that excessive regulation could stifle a rapidly evolving sector.

The document repeatedly stresses the need to “remove barriers to innovation” and expand access to data and testing environments for developers.

The recommendations also tie the regulatory strategy to broader economic and geopolitical goals, including maintaining American dominance in AI and preventing foreign competitors from gaining an edge through more permissive policies.

Congress is expected to weigh the proposal as debates intensify over how and how much to regulate artificial intelligence, with states in recent years moving ahead with their own AI-related laws in the absence of a comprehensive federal framework.