The U.S. Has 1,200 AI Bills but No Shared Test for Good Policy

The United States has become a patchwork of state-by-state legislation, with federal policy lagging behind. Arvind Krishna, CEO of IBM, recently warned that too many regulations can be “terrible,” while too few could lead to undesirable outcomes.

A recent surge in AI legislation has produced over 1,200 bills and nearly 150 enacted laws. Policymakers are working without a clear understanding of what constitutes good policy, leaving the country vulnerable in a global contest where technological competitiveness is essential for economic prominence and national security.

The current debate revolves around a binary choice between sweeping regulation and unrestricted operation, with too little attention given to how proposals might conflict with existing law. This approach oversimplifies the complexity of AI policy, which cannot be reduced to simple on/off switches or yes/no answers.

At the state level, bills often attempt to regulate “AI” as a category, ignoring that many uses fall within existing consumer protection, civil rights, intellectual property, and data privacy law. States like Colorado and Utah have passed omnibus statutes with reservations, attaching sunset clauses and delayed effective dates that signal their drafters’ uncertainty.

Even these attempts at regulation are being rolled back or narrowed in favor of targeted transparency measures. The federal conversation is similarly stuck. President Trump’s executive order directed the Department of Justice to challenge state AI laws, while the 2026 National Defense Authorization Act excluded preemption language entirely.

The recent controversy over Anthropic’s disclosure of Mythos Preview has prompted the White House to consider an FDA-like pre-release vetting system for advanced AI models. Internationally, the EU is implementing the AI Act, and China is deploying frontier capability under state direction. This raises the stakes in the US policy debate.

With the line between commercial AI and national-security capability collapsing, the cost of incoherent US policy will only continue to rise. The lack of a shared test for good policy is a fundamental problem that policymakers at every level must address. Until they can determine which specific regulation addresses which gap and at what cost, the country will remain vulnerable in a global competition.

The international backdrop only adds to the urgency of this challenge. The EU’s AI Act and China’s state-directed approach are setting the stage for a global competition where the US is increasingly at risk of being left behind. Unless policymakers can break through the current gridlock, the country will continue down a path that is both uncertain and potentially disastrous.

The fate of the U.S. in this global competition depends on its ability to adopt a more informed and nuanced approach to AI regulation. Anything less risks perpetuating a recipe for disaster in a world where technological competitiveness underwrites economic prominence and national security.