Bill > H0792

This bill establishes new liability standards for developers and deployers of certain artificial intelligence (AI) systems, specifically defining "high-impact artificial intelligence systems" and "generative artificial intelligence systems" as "products" subject to these regulations. Developers are held liable if they fail to exercise reasonable care in the design, providing adequate instructions or warnings, or adhering to express warranties, provided these failures are a proximate cause of harm, and the plaintiff can prove the developer knew or should have known of the danger, accounted for foreseeable uses, and that a feasible alternative design existed. Developers can benefit from a "safe harbor" if they conduct documented testing, disclose risks, maintain an AI data sheet, and take specific precautions for products used by minors. Deployers, defined as those operating an AI system for internal or third-party use (excluding individuals and small businesses), are generally liable as developers only if they make material changes to the product or intentionally misuse it against an express warranty; however, they can also achieve a "safe harbor" by implementing a documented risk management policy. The bill supplements existing laws and allows for comparative negligence, with a provision for damages to be reduced if another party's violation of the developer liability section contributed to the harm. Violations are considered unfair and deceptive acts in commerce, with enforcement authority granted to the Attorney General, and the act is set to take effect on July 1, 2026.

Business and Industry

https://legislature.vermont.gov/bill/status/2026/H.792