Relates to the training and use of artificial intelligence frontier models; defines terms; establishes remedies for violations.
The New York RAISE Act (S. 6953-B) amends the General Business Law by adding Article 44-B to regulate the training and deployment of high-powered AI systems called 'frontier models.' A 'frontier model' is defined in Section 1420 as an AI model trained using more than 10^26 computational operations with a compute cost exceeding $100 million, or a distilled model derived from such a system costing more than $5 million. A 'large developer' is any person who has trained at least one frontier model and spent over $100 million in aggregate compute costs, excluding accredited colleges and universities engaged in academic research. Under Section 1421, before deploying a frontier model, large developers must: implement a written safety and security protocol; retain unredacted copies of that protocol for the duration of deployment plus five years; publish a redacted version and transmit it to the Attorney General and the Division of Homeland Security and Emergency Services; record and retain test results in sufficient detail for third-party replication; and implement safeguards against unreasonable risk of 'critical harm.' Large developers are prohibited from deploying a frontier model if doing so would create an unreasonable risk of critical harm—defined in Section 1420 as causing death or serious injury to 100 or more people or at least $1 billion in damages through CBRN weapons or autonomous AI criminal conduct. Annual protocol reviews are required, and safety incidents must be disclosed to state authorities within 72 hours of discovery. Section 1422 authorizes the Attorney General to bring civil actions with penalties up to $10 million for a first violation and $30 million for subsequent violations, but establishes no private right of action. Section 1424 limits the law's scope to frontier models developed, deployed, or operating in whole or in part in New York. The act takes effect 90 days after enactment.
Status history
Current status as of 2025-12-19
Enacted
2025-12-19
observed 2026-08-27
Impact areas
- Enterprise Adoption
- Quality Assurance
- Certifications & Standards
- AI Policy