# New York's RAISE Act 2.0 Would Add Quarterly AI Reporting and Audits

- URL: https://usa-24.com/new-york-raise-act-2-0/
- Publisher: USA 24
- Section: Artificial Intelligence (AI)
- Published: 2026-10-31T14:01:27.831Z
- Updated: 2026-10-31T14:01:27.842Z
- Topics: AI Regulation, New York Legislature, Frontier AI Safety, State AI Laws, AI Preemption

> State Sen. Andrew Gounardes and Assemblymember Jordan Wright introduced a bill that would add quarterly reporting, 72-hour capability-jump disclosures and state-licensed examiners to New York's frontier AI law.

## Key takeaways

- The RAISE Act 2.0, introduced October 29, 2026, would require large developers to report quarterly on automated AI research and how they use computing power.
- Developers would have to disclose any substantial jump in risk or in their models' ability to automate AI research within 72 hours.
- Covered developers would retain a state-licensed frontier examiner for annual compliance audits and catastrophic-risk assessments every 90 days.
- The bill is only a proposal, and its sponsors have not said when a vote will happen.
- Congress has not enacted a federal AI law, while the White House has asked Congress to preempt state AI laws deemed unduly burdensome.

New York lawmakers want large AI developers to report every quarter on how much of their own research is automated, and to alert the state within 72 hours of a sudden jump in risk. State Sen. Andrew Gounardes of Brooklyn and Assemblymember Jordan Wright of Harlem introduced the bill on Thursday, October 29, 2026. It would tighten New York's frontier-AI safety law, the state's first law of its kind.

The bill is a proposal, not law, so its near-term effect is limited. The sponsors have not said when a vote will happen, and no committee action appeared in the reporting reviewed. It still matters because it pushes state oversight of frontier labs further at the same moment Washington is weighing whether to take that power away.

## What the New York RAISE Act 2.0 would require

According to the [sponsors' press release](https://www.nysenate.gov/newsroom/press-releases/2026/andrew-gounardes/sen-gounardes-proposes-major-new-ai-safety-and), the bill adds incident disclosure requirements, bolsters enforcement and establishes independent safety oversight. Its core obligations fall on large developers. Each quarter, they would report how much of their AI research is automated, which the bill calls "recursive self-improvement," and how they use their computing power.

The sponsors' framing is more pointed than the paperwork alone. The press release says the bill would let the state "pace the AI frontier" by forcing developers to slow down if their models pose serious danger. That language signals that reporting is a means to an end, a path to intervention rather than disclosure alone.

## Why the 72-hour clock and quarterly reports matter

Quarterly reports give regulators a steady view of how automated research is progressing. The 72-hour rule is the faster track. Developers would have to report any substantial jump in risk, or in their models' ability to automate AI research, within 72 hours. A quarterly filing alone could arrive long after a sudden capability gain.

The 72-hour window is not new to New York's law. Under the original RAISE Act, large developers must report within 72 hours after determining that a "safety incident" has occurred, according to [Davis Wright Tremaine](https://www.dwt.com/blogs/artificial-intelligence-law-advisor/2025/12/new-york-raise-act-ai-safety-rules-developers). The 2.0 proposal adds a second trigger tied to capability jumps, which is harder to define than a safety incident. Governor Kathy Hochul signed the original law on December 19, 2025, and it [takes effect January 1, 2027](https://policywindow.org/wiki/ny-raise-act).

The original law also ties its duties to frontier AI frameworks and Department of Financial Services disclosure filings, and it carries attorney general penalties of $1 million and $3 million, according to [AI Compliance Atlas](https://aicomplianceatlas.com/law/new-york-raise-act).

## How the frontier examiner and DIGIT would work

The most significant change may be the independent examiner. Developers would have to retain a state-licensed "frontier examiner" to audit their compliance with their own safety frameworks each year. The examiner would also run broader safety evaluations and produce catastrophic-risk assessments every 90 days.

PYMNTS compared the model to bank supervision, in which outside auditors check governance and controls. The comparison is useful. Banks already operate under examiners who sit outside the institution and report findings to regulators. A similar structure for AI labs would move safety checks from a company's own disclosures toward verification by a third party.

The bill would also give the newly created Office of Digital Innovation, Governance, Integrity and Trust, known as DIGIT, power to limit the computing power that large developers devote to training new models or to automated AI research. The state says the aim is to reduce catastrophic risk. On paper, that is the most aggressive provision in the bill. The reporting so far does not explain how DIGIT would use it or how the office would operate, and that gap matters to any lab deciding how much to spend on its next training run.

## Which companies could be covered, and what is still unknown

Frontier developers operating in New York, including large US labs such as OpenAI, Anthropic, Google, Meta and xAI, would face quarterly reporting, outside audits and possible compute limits if the bill passes. Coverage depends on thresholds that have not been confirmed for the 2.0 text.

The earlier numbers conflict. One summary of the original law cites a threshold of more than $100 million in aggregate computing costs. Another describes more than $500 million in annual gross revenue, including affiliates, for large frontier developers. Those figures come from different versions of the law, so neither should be treated as the test for the 2.0 bill until the text confirms it.

The main open questions are:

- Whether the bill advances, and when any vote might happen.
- Which covered-company thresholds would apply to the 2.0 text.
- Whether DIGIT would ever use its compute-limit power, and how the office would run.

## Where the chatbot law fits in New York's 2026 push

The frontier bill arrived the same week Hochul signed a separate measure aimed at minors. Senate Bill S9051B passed 60-0 in the state Senate and governs generative AI companions that chat with users by text or voice. According to [Rochester First](https://www.rochesterfirst.com/new-york-state/hochul-signs-new-york-laws-regulating-ai-chatbots-for-minors-and-online-baby-monitors/), the list includes ChatGPT, Gemini, Meta AI, Snap's My AI, Character.AI and Replika. The governor's office announced the signing in its [release on protecting kids from AI chatbots](https://www.governor.ny.gov/news/governor-hochul-signs-nation-leading-legislation-protect-kids-ai-chatbots-and-ensure-baby).

The effective date is unsettled. [Yahoo News](https://www.yahoo.com/news/politics/articles/show-id-chatgpt-york-ai-184110911.html) reports January 1, 2027, while Rochester First reports June 1, 2027. Companies should check the signed text before planning around either date.

The chatbot rules sit alongside a wider wave of scrutiny over minors and AI. Texas has sued [OpenAI and Sam Altman over ChatGPT safety for minors](https://usa-24.com/texas-sues-openai-chatgpt-minors/), showing that state attorneys general are also acting on the same issue.

## How federal preemption could shape the outlook

New York is not acting alone. [TechTimes](https://www.techtimes.com/articles/329592/20261027/record-115-state-ai-laws-2026-illinois-sets-frontier-audit-mandate-congress-still-silent.htm) counted 115 state AI laws across 30 states in 2026, a nearly 58 percent surge from last year. Illinois has become the first state to mandate annual independent third-party audits of its largest AI developers, so New York's examiner model looks less like an outlier.

Washington is pulling the other way. On March 20, 2026, the White House released its National Policy Framework for Artificial Intelligence: Legislative Recommendations, which [calls on Congress to preempt state AI laws](https://www.morganlewis.com/pubs/2026/03/white-house-ai-framework-puts-federal-preemption-at-the-center-of-the-debate) deemed unduly burdensome or inconsistent with federal policy. Congress has not enacted a single federal AI law, despite more than 150 federal AI bills introduced across the past two sessions.

Bores has warned that tech lobbyists will push Congress after the November election to eliminate state power to regulate AI. If that effort succeeds, New York's examiners and compute limits would never take effect, and labs would gain a single national standard. If Congress stays quiet, New York and other states would remain places where AI safety duties carry penalties. The same state-federal tension is playing out in health care, where [FDA and CMS are working to head off a clash with states over AI prescribing](https://usa-24.com/fda-ai-prescribing-rules-states/).

## What to do now if your company has New York exposure

Nothing in the 2.0 bill binds companies yet, so no action is required. Companies with New York exposure should still prepare:

- Check whether your frontier safety framework and incident-reporting process already meet the original RAISE Act, which takes effect January 1, 2027.
- Measure compute spending and revenue against both threshold figures reported so far, since the 2.0 numbers are unconfirmed.
- Assign owners for quarterly reports on automated research and for 72-hour capability-jump disclosures.
- Ask counsel whether any chatbot product used by New York minors falls under S9051B, and calendar both reported effective dates.
- Track the bill through the New York Senate website and the sponsors' press releases.

The next signals will be whether the bill gets a committee hearing and whether Congress takes up preemption after the November election. Either could turn a proposal into a template that other states copy.

## Frequently asked questions

### Is the New York RAISE Act 2.0 law yet?

No. Sen. Andrew Gounardes and Assemblymember Jordan Wright introduced the bill on October 29, 2026, and the reporting reviewed shows no committee action or vote date. The original RAISE Act, signed by Governor Kathy Hochul on December 19, 2025, takes effect January 1, 2027.

### Does the New York AI law apply to ChatGPT?

The original RAISE Act targets large frontier developers, and the 2.0 coverage thresholds have not been confirmed. A separate chatbot law, S9051B, which Hochul signed on October 28, 2026, governs generative AI companions including ChatGPT, Gemini and Character.AI. Reports disagree on whether it takes effect January 1, 2027 or June 1, 2027.

### What is a frontier examiner under the bill?

A frontier examiner would be an independent, state-licensed outside auditor hired by a covered developer. The examiner would conduct annual audits of compliance with the developer's safety frameworks. It would also run broader safety evaluations and catastrophic-risk assessments every 90 days.

### Could Congress override New York's AI rules?

The White House's March 20, 2026 framework calls on Congress to preempt state AI laws deemed unduly burdensome. Congress has not passed a federal AI law despite more than 150 bills, and no preemption bill has passed. Assemblymember Alex Bores has warned that tech lobbyists will push for preemption after the November election.

## Sources

1. [New York Bill Would Raise AI Safety Regulations - PYMNTS](https://www.pymnts.com/news/artificial-intelligence/2026/new-york-bill-would-raise-ai-safety-regulations/)
2. [Sen. Gounardes Proposes Major New AI Safety and Transparency Regulations - NY State Senate press release](https://www.nysenate.gov/newsroom/press-releases/2026/andrew-gounardes/sen-gounardes-proposes-major-new-ai-safety-and)
3. [New York RAISE Act: Responsible AI Safety and Education Act - Policy Window wiki](https://policywindow.org/wiki/ny-raise-act)
4. [New York RAISE Act: Frontier AI Law, Effective 2027 - AI Compliance Atlas](https://aicomplianceatlas.com/law/new-york-raise-act)
5. [Governor Hochul Signs Nation-Leading Legislation to Protect Kids From AI Chatbots - NY Governor's office](https://www.governor.ny.gov/news/governor-hochul-signs-nation-leading-legislation-protect-kids-ai-chatbots-and-ensure-baby)
6. [Show Your ID to Use ChatGPT? How New York's New AI Law Ends Online Anonymity - Yahoo News](https://www.yahoo.com/news/politics/articles/show-id-chatgpt-york-ai-184110911.html)
7. [New York enacts laws to ban AI companions for kids and eavesdropping baby monitors - Rochester First](https://www.rochesterfirst.com/new-york-state/hochul-signs-new-york-laws-regulating-ai-chatbots-for-minors-and-online-baby-monitors/)
8. [Record 115 State AI Laws in 2026: Illinois Sets Frontier Audit Mandate, Congress Still Silent - TechTimes](https://www.techtimes.com/articles/329592/20261027/record-115-state-ai-laws-2026-illinois-sets-frontier-audit-mandate-congress-still-silent.htm)
9. [White House AI Framework Puts Federal Preemption at the Center of the Debate - Morgan Lewis](https://www.morganlewis.com/pubs/2026/03/white-house-ai-framework-puts-federal-preemption-at-the-center-of-the-debate)
10. [New York Enacts RAISE Act for AI Transparency Amid Federal Preemption Debate - Davis Wright Tremaine](https://www.dwt.com/blogs/artificial-intelligence-law-advisor/2025/12/new-york-raise-act-ai-safety-rules-developers)