New York lawmakers have introduced a bill that would require large AI developers to face independent safety audits and report on how much of their research is automated. It would also give a state office power to limit computing devoted to AI development, putting oversight of AI systems on the legislative agenda in unusually concrete terms.
Watch Desk analysis
What happened
State Senator Andrew Gounardes and Assemblymember Jordan Wright introduced the bill on 29 October, according to PYMNTS. It would require large developers to report quarterly on automated research and their use of computing power, and disclose within 72 hours any substantial rise in risk or in a model’s ability to automate AI research.
The proposal would also require developers to retain an independent, state-licensed examiner for annual audits and broader safety evaluations every 90 days. The state’s DIGIT office could limit computing used for training or automated AI research, while the Attorney General could ask a court to halt dangerous activity, including a model’s deployment in New York. These are proposed powers, not rules already in force.
Why it matters
The bill would go beyond asking developers to publish their own safety plans. Independent examinations, regular assessments and short deadlines for reporting risk would give regulators more ways to scrutinise what companies are doing, and when.
The proposed computing limits are the sharper edge: oversight could reach beyond disclosure to the resources used to develop models. That is a substantial shift, and one that will put questions of authority and enforcement right alongside the bill’s safety aims.
Our read
This is a consequential proposal, not a new law. The audit and reporting requirements are specific enough to make the debate practical: who examines a developer, what counts as a serious change in risk, and how far should a state be able to restrict AI research? Washington’s approach to AI regulation is part of the backdrop, but New York’s next steps matter on their own.
What to watch
- Whether the bill advances through the New York legislature.
- How the proposal defines which developers count as large.
- What standards and powers would govern independent examiners and DIGIT.
Discussion spark: Should a state be able to restrict computing used for AI development when regulators judge the risk too high, or should that power remain with developers and courts?
Sources and evidence
- Source update (30 October 2026, 15:43 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.