New York 2025-2026 Regular Session

New York Senate Bill S06953

Introduced
3/27/25  

Caption

Relates to the training and use of artificial intelligence frontier models; defines terms; establishes remedies for violations.

Impact

The proposed legislation significantly aims to enhance accountability among large developers who engaged in the creation or deployment of frontier AI models. Under this act, developers are required to follow specific safety protocols designed to mitigate risks associated with their AI technologies. This effort seeks to not only prevent harm from AI deployment but also to reinforce trust in the AI sector by ensuring that developers are held accountable for their products. Failure to comply can result in severe penalties ranging from both civil fines to injunctive relief, reinforcing a culture of safety and responsibility in an industry that is rapidly evolving.

Summary

S06953, also known as the Responsible AI Safety and Education Act (RAISE Act), aims to establish comprehensive regulations surrounding the training and deployment of frontier artificial intelligence models in New York State. The bill introduces a new section to the General Business Law that mandates stringent transparency requirements for developers. They are required to implement written safety and security protocols, conduct annual independent audits, and disclose any critical safety incidents within a defined time frame. This bill seeks to prevent potential risks linked to advanced AI technologies that may operate with minimal human intervention.

Contention

Notably, the bill has raised discussions about the balance between innovation and safety. Supporters argue that regulating AI development is essential to prevent potential harms that unchecked technological advancement can bring, emphasizing the importance of safeguarding public welfare. On the other hand, some industry stakeholders express concern that overly stringent regulations may stifle innovation and impede the competitive advantage of New York developers in a global context. This duality of perspectives highlights the need for ongoing dialogue as AI technologies continue to unfold and integrate into various aspects of society.

Companion Bills

NY A06453

Same As Relates to the training and use of artificial intelligence frontier models; defines terms; establishes remedies for violations.

Similar Bills

CA SB1047

Safe and Secure Innovation for Frontier Artificial Intelligence Models Act.

NY A06453

Relates to the training and use of artificial intelligence frontier models; defines terms; establishes remedies for violations.

US SB5616

A bill to establish the Artificial Intelligence Safety Review Office in the Department of Commerce, and for other purposes.

CA SB294

Health care coverage: independent medical review.

CA SB53

CalCompute: foundation models: whistleblowers.

US SB4664

Department of Energy AI Act

NJ AR158

Urges generative artificial intelligence companies to make voluntary commitments regarding employee whistleblower protections.

NJ AR141

Urges generative artificial intelligence and content sharing platforms to make voluntary commitments to prevent and remove harmful content.