New York Creates Dedicated Unit to Enforce AI Safety Law
New York Attorney General Letitia James announced Tuesday the creation of a specialized enforcement unit to oversee the state’s Responsible AI Safety Education (RAISE) Act, a pioneering state law imposing mandatory safety, transparency and incident-reporting requirements on developers of the most powerful AI models. The AI Safety, Standards, Enforcement, and Transparency (ASSET) Unit will operate within the Attorney General’s Economic Justice Division and begin work ahead of the RAISE Act’s January 1, 2027 effective date [1][2].
What the RAISE Act Requires
The RAISE Act, signed into law on December 19, 2025, targets “large developers” of “frontier models” — defined as AI systems trained with more than 10^26 computational operations and a compute cost exceeding $100 million, or models derived from such systems through knowledge distillation [4]. A “large developer” is any entity that has trained at least one frontier model costing over $5 million and spent more than $100 million in aggregate on frontier-model training; accredited colleges and universities conducting academic research are excluded [4].
Before deploying a frontier model, covered developers must:
- Implement a written safety and security protocol specifying protections against “critical harm” — defined as death or serious injury to 100 or more people, or at least $1 billion in damages, caused or materially enabled by the model [4].
- Publish a redacted version of that protocol and transmit it to the Division of Homeland Security and Emergency Services; grant the Division or the Attorney General access to the unredacted protocol upon request, with redactions only as required by federal law [4].
- Retain detailed testing records sufficient for third-party replication [4].
- Undergo an annual independent third-party audit of compliance; publish a redacted audit report and transmit it to the Division of Homeland Security and Emergency Services; grant the Division or the Attorney General access to the unredacted report upon request, with redactions only as required by federal law [4].
- Report each “safety incident” — including autonomous model behavior, theft or escape of model weights, critical control failures, or unauthorized use — to the Division of Homeland Security and Emergency Services within 72 hours [4].
- Refrain from deploying a frontier model if doing so would create an unreasonable risk of critical harm [4].
The law also bars retaliation against employees who disclose safety concerns to the developer or the Attorney General [4]. Civil penalties reach 5% of total frontier-model compute costs for a first violation (15% for subsequent violations) and up to $10,000 per employee for whistleblower retaliation [4].
ASSET Unit Mandate and Structure
The ASSET Unit will lead enforcement of the RAISE Act and coordinate implementation of other New York AI laws, including:
- The SHIELD Act (data security requirements) [1].
- A chatbot safety law taking effect June 2027 banning unsafe features for children [1][3].
- General Business Law Article 47 (suicidal-ideation detection and AI disclosure for chatbots) [1].
- Penal Law §245.15 (criminalizing nonconsensual intimate imagery, including AI-generated) [1].
- The Synthetic Performer Disclosure Act (AI use in advertising) [1].
Beyond enforcement, the unit will develop policy proposals and provide technical expertise to other divisions within the Attorney General’s office [1][2]. It will work closely with the existing Bureau of Internet and Technology (BIT) [1][3]. Cristy Phillips, formerly Special Counsel for the Economic Justice Division, will lead the unit [1][2][3]. The office said it is “currently building the ASSET Unit team” [1].
A Pattern of State-Level AI Action
New York’s move reflects a broader trend. A November report from the National Association of Chief Information Officers found that at least 46 states have identified someone responsible for directing AI initiatives, but 56% lack dedicated AI budgets and 75% cite insufficient staffing as a major challenge [2]. California Governor Gavin Newsom recently ordered accelerated development of an independent AI safety auditing system [2]. New York’s approach adds a different dimension by giving the attorney general’s office a dedicated team focused on enforcing laws and investigating potential misconduct by technology companies [2].
Federal-State Tensions
The RAISE Act and ASSET Unit arrive amid unresolved disputes over state authority to regulate AI. In 2025, President Trump issued an executive order barring states from enforcing their own AI laws, sparking widespread backlash from lawmakers and advocacy groups [2]. Separately, James led a bipartisan coalition urging Congress to reject federal legislation that would preempt state regulations [1]. In June 2026, a coalition of state lawmakers warned against a federal proposal — the “Great American AI Act” discussion draft — that would authorize a federal AI standards center and preempt state laws, submitting a letter to Congress urging opposition [2]. James has consistently argued that states must retain the ability to act where federal legislation stalls [1][2].
What Remains Unknown
Several practical questions are unresolved. The compute thresholds — 10^26 operations and $100 million — capture only the largest current models; it is unclear how many companies will qualify as “large developers” when the law takes effect. The RAISE Act’s effective date is listed as January 1, 2027 in the Attorney General’s announcement [1], though the bill text states it takes effect 90 days after becoming law (signed December 19, 2025) [4]; the reason for the discrepancy is not explained in the announcement or bill text. How the unit will verify safety protocols, assess “unreasonable risk,” and coordinate with federal agencies also remains to be seen.
Sources
- Attorney General James Announces Creation of AI Safety, Standards ... (ag.ny.gov)
- New York attorney general creates unit to oversee AI safety and enforcement (statescoop.com)
- NY Attorney General James announces new unit to oversee AI ... (cbs6albany.com)
- NY State Senate Bill 2025-S6953A (nysenate.gov)