2025 Legislative Actions on AI
In 2025, state governments across the US began to adopt or evaluate AI tools to enhance public services. These efforts included improving customer service operations, strengthening inspections for health care facilities, and increasing roadway safety. At the same time, legislators, industry leaders, and other interested groups engaged in extensive discussions about the potential for AI misuse and the unintended consequences that may accompany its use.
During the 2025 legislative cycle, every U.S. state—along with Puerto Rico, the Virgin Islands, and Washington, D.C.—introduced at least one bill related to AI. Thirty‑eight states ultimately passed or enacted approximately 100 measures. Illustrative examples include:
Arkansas passed a law clarifying ownership rights for AI‑generated material. Under this legislation, the individual who provides the input or data used to generate the content—or an employer, if the content is produced as part of an employee’s job duties—holds ownership. The law also affirms that such generated material must not violate existing copyright or intellectual property protections.
Montana enacted a “Right to Compute” statute that establishes expectations for critical infrastructure systems governed by AI technology. The law requires system deployers to adopt a risk‑management plan informed by designated standards, including the most recent NIST AI Risk Management Framework. It also limits the government’s ability to restrict the private ownership or use of computing resources for lawful purposes unless such restrictions meet a compelling state interest.
New Jersey adopted a resolution calling on companies that develop generative AI systems to voluntarily commit to protecting employees who report misconduct or policy violations.
New York passed a law requiring state agencies to publicly disclose detailed information about the automated decision‑making systems they use by maintaining an inventory overseen by the Office of Information Technology. The law also makes changes to civil service protections to ensure that government use of AI does not diminish employee rights under collective bargaining agreements or lead to job displacement.
North Dakota expanded its stalking and harassment statutes to bar individuals from using AI‑enabled robots to track, intimidate, or harass others.
Oregon approved legislation stating that AI systems or other non‑human agents may not use professional titles reserved for licensed or certified medical personnel, such as “registered nurse” or “certified medication aide.”