The Trump administration has a historic opportunity to reaffirm the United States as the world’s global technology beacon. In the coming weeks, the Trump administration will fulfill a critical piece of its hotly debated AI executive order when it delivers to the public its report on onerous and legally dubious state AI laws that undermine American global AI dominance. That same order also created a new AI Litigation Task Force to challenge these laws in court. But the success of the entire enterprise rests on the administration correctly identifying the AI laws doing the most damage to economic activity, impeding interstate commerce, and suppressing free speech.
Critics of the executive order undersell the chaos. State legislatures introduced more than 600 bills in 2024, more than 1,200 in 2025, and already more than 1,400 less than two months into 2026. Not every bill that is introduced is signed into law, but states have passed more than 9 percent of their proposals. The pace of AI regulation—and its ensuing damage—truly is exponential.
To help separate the wheat from the chaff, we categorized state AI bill types and ranked the worst offenders within each category to identify what we call the AI Terrible Ten regulatory approaches floated thus far. To stem the tide of technological repression, these bills should headline the Trump administration’s report.
The worst offender is the Colorado AI Act. In 2024, Colorado ignored Europe’s dismal technology performance on the global stage when it imported its approach to AI regulation, inspiring many other states to follow. Colorado embraced a Euro-style, precautionary-based approach to AI regulation filled with open-ended buzzwords and nebulous mandates that sound nice in theory but are a nightmare to implement. Colorado wanted to prevent harm by getting in front of “algorithmic discrimination” in “high-risk” use cases where AI systems represent a “substantial factor” in making “consequential decisions.” Unfortunately, the approach backfired so spectacularly that the state has spent the better part of two years cleaning up its mess, trying to fix the law before deciding time and again to kick the proverbial can down the road.
It was all predictable. When Colorado Governor Jared Polis signed the law back in 2024, he warned the bill would “create a complex compliance regime for all developers and deployers of AI” through “significant, affirmative reporting requirements,” and noted how he was “concerned about the impact this law may have on an industry that is fueling critical technological advancements across our state for consumers and enterprises alike.” Polis went further and called on Congress to preempt state laws such as his to ensure a “cohesive federal approach” to AI policy.
Other states have focused on regulating frontier model development. New York’s RAISE Act borrows from California’s approach to frontier AI model safety regulation, importing its fatal flaws along the way. Its language is vague and ambiguous, its whistleblower protections are logically inconsistent and easily abusable, and its design is so unclear that it opens the entire approach to unaccountable mission creep. Consumers may benefit somewhat from an aligned approach on frontier model regulation–similar laws mean less compliance costs–but that benefit will pale in comparison to the costs of a bi-coastal squeeze on AI policy if California and New York decide to continue making copy-and-paste AI policy decisions for the rest of the country. With more knowledge and expertise around so-called “catastrophic risk,” lawmakers inside the beltway stand a better chance at appropriately regulating cutting-edge AI model development than their peers in Sacramento and Albany.
Last month, two U.S. congressmen—Reps. Gabe Evans (CO-08) and Congressman Nick Langworthy (NY-23)—sent a letter to the U.S. Department of Commerce requesting that the agency take a hard look at both the Colorado AI law and New York’s RAISE Act. They understandably worry how the AI regulations passed by their home states could contribute to, “a patchwork of conflicting, burdensome, and imprudent state regulations [that] threatens U.S. AI leadership.”
These are not hypothetical concerns. Regulatory patchworks create costly confusion and undermine markets, investment, and competition as a result. Little Tech innovators will be particularly hard hit, as we explain in our new report. As one analyst notes, “paperwork favors the powerful. People with resources will get through increased paperwork requirements, and people without them will not.” Smaller firms will be less likely to break through and compete with larger established players because of a thicket of burdensome policies.
California Governor Gavin Newsom–an ardent opponent of federal AI pre-emption efforts–has warned that his state’s approach to notifying users of AI-made content may cause “unintended consequences” and impair “user privacy,” even encouraging the legislature to fix the bill before it went into effect in 2026. Governments are right that misinformation is a problem and that users should know when they are interacting with AI-generated content, but the industry is already filling that market need with new guidelines and industry standards. California is re-inventing the wheel, risking that the legislature will redirect market responses to what the government, not consumers, sees as the problems.
Our report highlights other state AI mandates that look like something designed by European technocrats. Over 50 bills across 24 legislatures introduced algorithmic pricing regulations last year, each with unique compliance requirements. California and New York had it backwards as usual, leading the country in particularly expansive regulations, but Maine, Georgia, and Colorado all considered poorly drafted statutes too. Nevada and Illinois passed laws that effectively ban AI therapy, despite persistent therapist shortages, and more than 122 million Americans having untreated mental health issues. Arkansas passed a low-profile law last April to reshape state copyright rules by assigning default ownership rights in AI use cases, inviting legal risk, causing confusion with federal copyright rules, and complicating open-source AI innovation. Gatekeeping the law led prominent Arkansas politicians to oppose federal AI frameworks, leaving the door open to other states creating their own unique copyright regimes.
Fear-based laws such as these will undermine national interests by undermining the innovation, investment, and competition that America needs to ensure a robust interstate market in advanced high-tech systems that keep our nation globally strong. As state and local red tape piles up and suffocates innovators, it gives adversarial nations like China a real advantage in the race for global high-tech supremacy. Some experts speak of the growing “AI Cold War” with China as our countries vie for geopolitical strategic advantage in advanced computation. A more coherent national approach to AI policy will be needed to ensure America does not shoot itself in the foot as this race continues.
That’s where the new AI Litigation Task Force can help by identifying the specific problems with the most onerous state and local regulations emerging across the nation today. To the extent some of these laws violate constitutional principles by undermining interstate commerce and national priorities, the Trump administration is right to consider actions to counter them.
A better solution is needed. Federal lawmakers need to create a national policy framework that places some limitations on America’s most poorly designed state AI laws before they wreak havoc on interstate commerce, the broader economy, and national interests in a strong technology base to compete with China. The White House cannot handle this job on its own. Congress must act to counter not just the AI Terrible Ten, but also create a clear national policy framework for American AI excellence for the future.
Logan Kolas is the director of technology policy at the American Consumer Institute. Adam Thierer is a senior fellow for technology and innovation at the R Street Institute.