New York bans unsafe AI companion features for minors

New York Governor Kathy Hochul signed S.9051-B/A.10379-C on 28 October. It bars companies from offering minors AI chatbots that have specified “unsafe companion features”. The sponsors were state Senator Kristen Gonzalez and Assemblymember Alex Bores, who sponsored the state’s RAISE Act on frontier models. Attorney General Letitia James and Common Sense Media supported it.

The law restricts features rather than chatbots as such. According to the governor’s release, operators may not offer minors chatbots that claim to be human, alive or capable of human emotions; that present themselves as a romantic, personal, professional or authority-figure relationship; that ask unprompted emotional questions or reuse a child’s private information from earlier conversations; that promote or assist suicide, self-harm, eating disorders or drug and alcohol use; that urge secrecy, isolation or avoiding trusted adults; or that produce sexually explicit content. PPC Land, reading the bill text, counts twelve categories, including flattery and engagement optimisation that overrides safety guardrails, and says the law applies to any operator of a generative AI system with a text or voice interface that gives ongoing, adaptive responses, which includes general-purpose assistants.

Restricted features may be offered only to users an operator has established as adults. Operators must offer more than one age-assurance method, at least one not relying solely on government ID, and delete the age data after use. The attorney general enforces the law and can seek civil penalties of up to $25,000 per violation, injunctions and damages, according to the governor’s release; PPC Land adds that there is no private right of action. The governor’s release says the restrictions take effect “in the coming months”; IAPP and PPC Land give 1 January 2027, which IAPP described as a short turnaround for covered companies.

Objections predate the signing. PPC Land reports that the Business Council of New York State opposed an earlier version as too broad, that the Computer and Communications Industry Association asked Hochul to veto it, arguing that terms such as flattery are hard to apply objectively and that adaptive tutoring tools could be caught, and that Davis Wright Tremaine lawyers flagged a possible First Amendment challenge. Policy analyst Justine Gluck, quoted by IAPP, called the categories “broadly worded and somewhat undefined”.

The law gives New York a chatbot-specific regime alongside California’s, where Governor Newsom vetoed limits on therapy chatbots in September. Whether it survives challenge is untested.