New York wants independent auditors for frontier AI: the bill that turns safety into bank-style supervision
Direct answer
State Senator Andrew Gounardes of Brooklyn and Assemblymember Jordan Wright of Harlem introduced bill S.10731 on October 29, 2026, expanding the RAISE (Responsible AI Safety and Education) Act passed in 2025. The current RAISE Act requires large developers to publish safety frameworks and report critical safety incidents. The new bill adds what the current law lacks: mandatory independent auditing, with a state-licensed frontier examiner reviewing compliance every year and assessing catastrophic risk every 90 days. The bill is compared to bank supervision, where outside auditors evaluate governance, operational risk, controls and remediation. It is an introduced bill, not current law.
The five changes it proposes
- Independent frontier examiner: each large developer must retain a state-licensed examiner to audit compliance with its safety frameworks every year, run broader safety evaluations and assess catastrophic risk every 90 days.
- Quarterly reports on research automation: developers must report each quarter how much of their AI research is automated, known as recursive self-improvement, and how they use their computing power.
- 72-hour disclosure: any substantial jump in risk or in models' ability to automate AI research must be disclosed within 72 hours.
- DIGIT can limit compute: the new Office of Digital Innovation, Governance, Integrity and Trust (DIGIT) could limit the computing power large developers devote to training new models or to automated AI research, to reduce the risk of catastrophic harm.
- Fines and personal certification: penalties rise to as much as $1 million per day for a first violation and $3 million per day for later ones, and a senior officer must certify filings under penalty of perjury. The bill also reinforces that state labor-law whistleblower protections apply to large developers' employees and bans contracts that silence them.
What it means for your business
- If you buy frontier AI, ask for audit evidence: the bill's direction is clear: publishing a safety plan is no longer enough, and independent auditing becomes the expected standard. At your next renewal with a large provider, ask for its published safety framework, its reported incident history and, when available, its independent audit.
- One year after RAISE: the bill arrives one year after the original law and after incidents in which models put safety and the economy at risk, per Gounardes' statement. The sequence matters: transparency was required first, verification is now proposed.
- The contrast with the chatbot law: on October 28, New York signed a law banning specific features for minors. On October 29, it introduced a bill supervising frontier developers. One front regulates the product that touches the user; the other regulates the lab that builds it.
Want AI that works for your business?
Message us on WhatsApp and let's talk about automation and AI agents for your business.
WhatsApp inicIARelated articles
- New York signed its chatbot law for minors: 12 features banned from January 1, 2027
- Connecticut became the strictest state for business AI on October 1
- How to choose an AI vendor: 7 questions before signing
Frequently asked questions
What is New York bill S.10731?
A bill introduced on October 29, 2026 by Senator Gounardes and Assemblymember Wright expanding the 2025 RAISE Act with mandatory independent auditing for large AI developers.
Is it already law?
No. It is an introduced bill. The current law is the 2025 RAISE Act, which requires publishing safety frameworks and reporting critical incidents.
What does a frontier examiner do?
It audits the developer's safety-framework compliance every year, runs broader safety evaluations and assesses catastrophic risk every 90 days, under a state license.
What fines does it propose?
Up to $1 million per day for a first violation and $3 million per day for later violations, with senior-officer certification under penalty of perjury.
