Assembly Passes RAISE Act to Put Vital Safety Standards in Place for Artificial Intelligence Frontier Models

Speaker Carl Heastie and Assemblymember Alex Bores today announced that the Assembly has passed the Responsible Artificial Intelligence Safety and Education (RAISE) Act, which will put critical safety standards in place for artificial intelligence (AI) frontier models, including creating a safety plan to prevent severe risks, having the safety plan reviewed by a third party and disclosing major security incidents (A.6453-B, Bores).

“As technology changes and evolves, it is critical that our laws adapt with them,” Speaker Heastie said. “With the rate at which AI is advancing and the extent of what these models are capable of producing, to have a safety plan in place is common sense.”

“We need to take heed when the leaders in this field say that there need to be safety precautions taken with this rapidly developing technology,” Assemblymember Bores said. “The RAISE Act will ensure there are protocols in place in the event of a major security risk and help New York to react and prevent future incidents.”

In March 2023, more than 1,000 tech leaders called for a pause for at least six months on training frontier models until international safety standards could be established. While the pause did not occur, AI models have become more powerful and have shown the ability to be used by bad actors to for devastating purposes.

The RAISE Act would require that AI developers put into place a safety plan for their frontier models to prevent severe risks. The safety plan would then need to be reviewed by a qualified third party. Additionally, they would be required to disclose major security incidents to the attorney general and the Division of Homeland Security and Emergency Services to prevent similar incidents from occurring in the future.