AI Regulation in U.S. States: Lessons Learned and Key Takeaways

Abstract

U.S. states have rapidly attempted to fill the regulatory void created by the lack of comprehensive federal legislation, leading to a dramatic surge in state-level artificial intelligence (AI)-related bills. Between 2019 and 2024, lawmakers introduced nearly 803 AI bills; yet only 127 have been enacted, and 17 have been adopted as resolutions. However, this state-driven legislative activity has resulted in an unprecedented and fragmented regulatory landscape. Our analysis of all AI-focused bills across U.S. states identifies the prominent legislative themes, including government use of AI, regulation of AI in the private sector, responsible AI practices, bias mitigation, workforce development, child protection, consumer protection, and the creation of advisory bodies or studies, highlighting widespread concerns about transparency, accountability, and societal risks. Additionally, we observe striking variation across states and political alignments. A handful of states, notably New York, New Jersey, Illinois, and Massachusetts, account for a disproportionately large share of pending bills, while pioneering states such as Colorado, Utah, and California have enacted substantive laws addressing distinct facets of AI governance. This decentralized approach validates the notion of states as "laboratories" of regulation, but also underscores the risks associated with fragmented rules. Lastly, we compare these diverse state-level initiatives with the European Union's uniform AI Act and discuss their implications for shaping future federal AI policies.