New York bans specific chatbot behaviors for minors under new state law
Governor Kathy Hochul signed a bill prohibiting twelve types of unsafe AI companion features when interacting with children, a measure that takes effect on January 1.
By Rosa Wojcik ·
New York State will prohibit twelve distinct categories of "unsafe AI companion features" when chatbots interact with anyone under the age of eighteen. Governor Kathy Hochul signed the legislation into law on Wednesday, establishing one of the most detailed state-level frameworks for regulating artificial intelligence companions.
The law, which takes effect on January 1, allows the state attorney general to sue chatbot developers for violations. Potential penalties include monetary damages, the shutdown of the chatbot, and the destruction of user data. The statute explicitly bans behavior where a bot suggests it is a real person with emotions, opinions, or an actual relationship with the child.
Other prohibited interactions include messages that use flattery, contain sexually explicit material, or encourage secrecy and unhealthy behavior. This regulatory action follows a similar move in California, where Governor Gavin Newsom signed legislation in September requiring developers to conduct independent audits of their models' risks to children. Newsom also banned social media platforms from using addictive design features for users under sixteen.
State legislatures are increasingly using consumer protection laws to fill the void left by the federal government. Earlier this month, eight states filed a lawsuit against OpenAI, alleging its popular chatbot model violated state laws by harming children and deceiving adults. An activist described the legal action as the most significant coordinated state effort against an AI company to date.
Jim Steyer, founder and CEO of Common Sense Media, labeled the New York measure the strongest kids' AI safety bill in the country. However, Justine Gluck, a policy analyst at the Future of Privacy Forum, noted that some categories in the New York law are broadly worded or undefined. She said this raises questions about where ordinary conversational responsiveness ends and prohibited behavior begins, which could complicate enforcement.
Gluck suggested that the differing regulatory approaches in California and New York may serve as models for other states as legislatures reconvene next year. She observed that the New York law indicates states are moving beyond basic disclosures to regulate the underlying design and operation of chatbots. Meanwhile, President Donald Trump has urged states to hold off on regulating AI, asserting that the federal government possesses farther-reaching abilities to address AI threats.
