The New York Attorney General has created a specialized unit to investigate artificial-intelligence risks, enforce state technology laws and give technical support to cases across the office. The AI Safety, Standards, Enforcement and Transparency Unit, known as ASSET, arrives as New York prepares to implement new rules for frontier-model developers and consumer-facing chatbots. Its creation consolidates expertise, but it does not create powers beyond the statutes the office can enforce.
ASSET will lead enforcement of the Responsible AI Safety and Education Act when that law takes effect January 1, 2027. The attorney general says the law requires covered developers to disclose safety measures publicly and report specified safety or security incidents quickly. The scope is aimed at large frontier-model developers, not every business that uses a spreadsheet, chatbot or automated recommendation. Coverage depends on the law’s definitions and any implementing guidance.
The unit will also work on privacy, cybersecurity, digital scams and online safety. Those subjects overlap with existing laws rather than fitting into a single AI category. The SHIELD Act requires reasonable data-security practices, consumer-protection law addresses deception and abuse, and criminal law covers some harmful distribution of intimate images, including altered or synthetic material. ASSET’s technical role may help investigators connect a system’s operation to the correct legal theory.
Other state requirements are arriving on different schedules. New York has a law requiring chatbot disclosures and reasonable efforts to detect and address suicidal ideation, and legislation restricting unsafe chatbot features for children is set to take effect in June 2027. The attorney general’s announcement also cites advertising disclosures for synthetic performers. Businesses should not treat these measures as one interchangeable compliance rule; each has its own covered conduct and effective date.
ASSET is being placed in the Economic Justice Division and will work closely with the Bureau of Internet and Technology. Cristy Phillips, previously special counsel for the division, will lead the unit, and the office says it is building the team. The announcement does not state a budget, staffing count or enforcement calendar. Those details will matter in judging whether the unit adds investigative capacity or mainly coordinates work already distributed across bureaus.
A technical unit can improve cases involving model testing, data flows, security controls and claims about automated products. It can also reduce the risk of describing ordinary software as an advanced AI system. Investigators will need repeatable methods for preserving evidence, evaluating vendor assertions and separating a model’s design from the way a customer deploys it. Public guidance can make those methods more predictable without revealing confidential investigative information.
For companies, the immediate task is an inventory of systems, vendors and legal obligations. Developers covered by the RAISE Act need incident-reporting and disclosure processes. Organizations using third-party tools should know what data they send, which party controls updates and how a harmful output is escalated. A contract label that calls a product a service does not replace an analysis of the organization’s own responsibilities under privacy, security and consumer laws.
For the public, the unit creates a clearer destination for complaints and whistleblower information, though it is not a substitute for emergency help or private legal advice. The office should publish accessible explanations of which reports belong with ASSET, what information is useful and how confidential material is handled. Aggregate statistics on inquiries, investigations and outcomes would help residents see whether the new structure produces practical protection.
The unit’s credibility will depend on technically sound, evenhanded enforcement. AI safety can involve serious risks, but broad claims unsupported by system evidence can mislead consumers and regulated businesses alike. New York now has a named team responsible for turning several laws into oversight. Its next tests are to define priorities, explain compliance expectations before effective dates and show that investigations distinguish demonstrable harm from speculation.
Primary sources: New York Attorney General ASSET Unit announcement; New York State Senate RAISE Act record. Reviewed November 17, 2026. Development: ASSET Unit announced November 17, 2026.
