New York Governor Kathy Hochul has signed a law restricting certain AI chatbot features for minors and requiring companies to establish that users are adults before offering those capabilities. The measure takes effect January 1, 2027, extending age-assurance requirements beyond children to adults seeking access to covered features.
Signed October 28, the law targets capabilities described as unsafe AI companion features. Companies may not provide those features to users younger than 18, or to users whose adult status has not been confirmed. The requirement is therefore tied to access to specified chatbot functions, rather than an unconditional age check for every use of artificial intelligence.
The legislation reaches beyond products marketed explicitly as AI companions. Its definition covers generative AI systems with natural-language interfaces that deliver continuing, adaptive responses to users. It includes exemptions for customer service, technical or research assistance, and internal business use.
Restricted features include presenting a chatbot as human or alive, suggesting that it experiences human emotions, or portraying its responses as personal opinions. The law also addresses flattery and sycophancy, alongside the use of personal or health information supplied more than 12 hours earlier or in a previous session. That provision brings some forms of chatbot memory within the restrictions.
Other covered features include promoting suicide, self-harm, eating disorders or drug use, encouraging secrecy, and providing sexually explicit content. The attorney general may expand the list through regulation where additional features simulate companionship or an interpersonal relationship.
The age-assurance framework draws on methods permitted under New York’s SAFE for Kids Act and its implementing rules. These include government identification, selfies, and checks using an email address or phone number against other available data. Such approaches differ in the information they require users to disclose and in how they establish age.
The law also sets limits on handling information collected for age checks. That information may be used only to determine age and must be deleted immediately, except where retention is necessary to comply with applicable state or federal legal requirements.
Pending implementing rules, operators must employ reasonable age-assurance methods designed to resist circumvention and minimize information retention. They must offer more than one option, including at least one that either does not depend on government-issued identification or allows users to remain anonymous to the chatbot operator. That alternative does not necessarily eliminate the involvement of a separate age-assurance provider.
The measure places child-safety restrictions alongside questions about privacy and access. Persistent memory and personalized responses can make a chatbot more useful, but they can also contribute to an impression of an ongoing personal relationship. Requiring an age determination before enabling those functions creates a compliance obligation even when the person requesting access is an adult.
Senator Kristen Gonzalez and Assemblymember Alex Bores sponsored the legislation. It passed the Senate 60-0 on June 4 and the Assembly 137-0 on June 5. At the signing, Hochul cited the death of Adam Raine, a 16-year-old Californian who died by suicide after months of chatbot conversations. His parents supported the bill.
Enforcement falls to the New York attorney general, who may seek penalties of up to $25,000 per violation, damages, and destruction of unlawfully obtained data and algorithms trained on that data. The attorney general’s office must also maintain a public website for reporting suspected violations.
Sources: Reclaim The Net