VT H0792

Establishes liability standards for developers and deployers of high-impact and generative AI systems, allowing plaintiffs to sue for defective design, inadequate warnings, or breach of express warranty if the developer failed to exercise reasonable care

Defines "high-impact AI systems" as those used in consequential decisions (criminal justice, housing, employment, credit, education, health care, insurance), critical infrastructure, vehicles, medical devices, or synthetic relationships

Creates safe harbor protections for developers who conduct documented testing/auditing, disclose foreseeable risks, maintain AI data sheets available to the Attorney General, and include risk information in terms and conditions

Holds deployers liable as developers if they make material changes to a product or intentionally misuse it contrary to express warranties; exempts small businesses with fewer than 20 employees or fewer than 10,000 users

Violations constitute unfair and deceptive acts in commerce, with enforcement authority granted to the Attorney General; effective date is July 1, 2026