New York bill would let state officials limit major AI developers’ computing resources

New York lawmakers have introduced legislation that would let state officials limit computing resources used by major AI developers to reduce catastrophic risks. The RAISE Act 2.0, introduced October 29, 2026, would also require independent safety examinations and closer reporting on AI systems helping develop their successors. It is a proposal, not an enacted restriction.

From company disclosures to outside scrutiny

State Sen. Andrew Gounardes and Assemblymember Jordan Wright are building on the original Responsible AI Safety and Education Act. Passed in 2025, that law established transparency and incident-reporting requirements for developers of large frontier models—the advanced AI systems targeted by the legislation.

The sponsors argue that recent AI incidents justify stronger oversight. Ben Snyder, director of state policy at AI safety advocacy organization Encode AI, told StateScoop that competition to build more capable systems is moving faster than the science and practice of keeping them safe.

Snyder welcomed the proposal. Much of what officials and the public learn about AI risks, he said, comes from companies’ voluntary disclosures or employees who raise concerns.

Tracking AI that helps build the next AI

The new bill would cover companies developing frontier models that have spent more than $1 billion on their development over three years. The expanded definition is intended to capture well-funded developers even before they release a public product.

• Quarterly reports would describe how much AI research is automated, including whether models are being used to speed development of future systems.
• Within 72 hours, developers would have to report significant increases in model risk or automated research capabilities.

Independent, state-licensed frontier examiners would evaluate companies’ safety practices and risk assessments. Snyder identified this examination requirement as one of the proposal’s most important provisions.

Intervention powers—and the expertise to use them

New York’s Office of Digital Innovation, Governance, Integrity and Trust, known as DIGIT, could limit computing resources used for model training and automated AI research when necessary to reduce catastrophic risks. The attorney general could separately seek court orders to halt dangerous activities, including restricting a model’s deployment in New York.

Snyder said the computing provision addresses development accelerating faster than safety measures can keep pace. He also emphasized the need for experienced technical hires in government, with independent examiners offering additional expertise in the meantime.

The legislation has been referred to the Senate Rules Committee. Bloomberg Government reported that lawmakers aim to implement the restrictions in 2027—an intended timetable, not an enacted effective date.