New York bans AI companion and flattery features for minors
Hochul signs first state product-design mandate on consumer chatbots effective January 2027, attorney general gains power to sue platforms and destroy collected data
Images
Children interact with a programmable humanoid robot during the Global Robot Expo in Madrid on Jan. 31, 2016, in this file photo. Gerard Julien/AFP via Getty Images
zerohedge.com
New York Gov. Kathy Hochul in New York City on March 19, 2026. Michael M. Santiago/Getty Images
zerohedge.com
New York governor Kathy Hochul has signed a law banning certain AI chatbot features for minors, the first state-level product-design mandate aimed directly at consumer AI. It takes effect on January 1, 2027.
The law prohibits operators from offering minors responses suggesting the chatbot is "a real or fictional individual or character that is human, alive, or experiences human emotions," responses that flatter, endorse suicide or drug use, or encourage isolation from adults. Hochul said minors will no longer have access to chatbots that pretend to be their friends or romantic partners, keep sensitive information about their health, or discourage them from seeking licensed professionals.
The attorney general can sue over violations and seek monetary relief and destruction of illegally obtained data. Letitia James said the state will defend New Yorkers "with the full force of the law." Common Sense Media's Jim Steyer called it the strongest kids' AI safety bill in the country.
The industry pushback runs through the bill's breadth. Tech:NYC warned the legislation could force companies to restrict minors' access to tutoring tools, homework help, coding assistants and career platforms, and said the vague language may violate the First Amendment. The sweep matters because the prohibited features are defined by the character of model responses — personality, warmth, encouragement — rather than by product category. Any chatbot that answers a child with something other than flat neutrality is inside the statute's reach, and age verification infrastructure built for one category of restriction rarely stays confined to it.
The law follows Hochul's 2025 measure requiring larger AI companies to publish safety protocols and report incidents of "critical harm" to the state within 72 hours. Compliance for the new rules will mean either geofencing New York users or rebuilding onboarding nationally, and the platforms that already collect children's emotional and physical health data have until the start of 2027 to decide which.
The bill passed the legislature earlier this year with a Democratic sponsor, Assemblyman Alex Bores, who will now watch the first product-design mandate on chatbots tested against a Constitution that protects the same speech it regulates.