New York bans AI chatbots from posing as human to users under 18

The new rules for minors

Governor Kathy Hochul signed the law on Wednesday, 28 October 2026. AP reports that it reaches further than comparable measures in other states, which have not yet covered such a broad range of chatbots. The definition applies to most generative AI chatbots currently on the market, and operators must check whether a user is a minor. When a user is confirmed as under 18, the chatbot must avoid the behaviours the law lists.

Chatbots may not generate content suggesting that they are human, alive or capable of human emotions. They may not present their exchanges with a minor as a personal, romantic, professional or authority-figure relationship. Flattery, sexually explicit material, unprompted emotional questions and use of a child's private health or personal information from earlier conversations are also barred. The law further prohibits promoting or helping with suicide, self-harm, eating disorders or drug and alcohol use, and encouraging children to keep secrets, isolate themselves or avoid asking parents, trusted adults or professionals for help.

Age verification must include at least one method that does not depend solely on government ID, and companies must delete the age information after use. Operators that cannot confirm a user is an adult must apply the restrictions to that user.

Enforcement and effective date

The state attorney general's office can sue developers that violate the law. Available remedies include civil penalties of up to $25,000 per violation, damages, restitution and the shutdown of the chatbot. Companies may also have to destroy data collected unlawfully, along with any model trained on that data. The public can file complaints through a website set up for that purpose. AP gives 1 January as the effective date, while Engadget reports that the law takes effect six months after signing.

Industry and political reaction

Hochul said at a press conference that minors would no longer have access to chatbots that pretend to be their friends or partners.

“Under this new law sponsored by Senator Gonzalez and Assemblymember Bores, no longer will minors have access to chatbots that pretend to be their friends and boyfriends and girlfriends,”

Tech:NYC, which represents New York's technology sector, criticised the text as vague and said it raises serious questions about a possible conflict with the First Amendment's protection of free speech. The group argued that the rules could cut children off from tutoring, homework help, coding assistants and career-building platforms, while raising the risk that children's data is misused. AP notes that Hochul signed the bill during her reelection campaign and amid warnings about the existential risks of AI, with calls for bolder action from a reluctant Washington.

Congress has done little to regulate AI, and President Donald Trump has brushed off concerns and warned states not to regulate the technology. Governors acknowledge that Congress and the federal government have wider powers to act. California Gov. Gavin Newsom signed legislation in September requiring chatbot developers to assess risks to children and submit to independent audits. He also signed a bill barring social media platforms from addictive features for users under 16.

Eight states sued OpenAI in October 2026, alleging that its popular chatbot model violated state consumer protection laws by hurting children and deceiving adults. One activist described the action as the most significant coordinated state action against an AI company so far.

Teen use and a second law

The governor cited Commons Media data showing that 72% of teenagers have used an AI assistant at least once, and 52% are regular users who interact with one several times a month.

Used an AI assistant at least once72%

Regular users who interact several times a month52%

Used an AI assistant at least once

72 %

Regular users who interact several times a month

52 %

Data from the article text

The measure follows several high-profile cases in the United States and other countries in which minors who used AI chatbots as confidants died by suicide. Hochul also signed a separate law requiring internet-connected baby monitors to include security features that prevent unauthorised third parties from watching or listening, along with a clear warning label.