Bill > A06453

This bill, known as the Responsible AI Safety and Education (RAISE) Act, establishes comprehensive regulations for the development and deployment of "frontier models" - advanced artificial intelligence systems with high computational costs. The bill defines a frontier model as an AI model trained using over 10^26 computational operations with a compute cost exceeding $100 million, and imposes significant transparency and safety requirements on "large developers" of such models. Key provisions include mandating that developers create detailed safety and security protocols, conduct annual independent audits, disclose safety incidents within 72 hours, and implement safeguards to prevent critical harm. The bill also provides strong whistleblower protections for employees who report potential risks, prohibiting retaliation and allowing them to seek legal remedies. Developers must publicly share redacted versions of their safety protocols and grant access to state agencies like the Division of Homeland Security and Emergency Services. Violations can result in substantial civil penalties up to $30 million, with the Attorney General empowered to bring enforcement actions. The legislation applies specifically to AI models developed or operated in New York State and aims to proactively address potential risks associated with advanced artificial intelligence technologies by establishing a comprehensive regulatory framework focused on safety, transparency, and responsible innovation.

Business and Industry, Housing and Urban Affairs

https://www.nysenate.gov/legislation/bills/2025/A6453