New York AI Law: State Bans 'Unsafe' Chatbot Traits for Under-18s

Governor Kathy Hochul signed a new state law on Wednesday, establishing strict rules for how artificial intelligence companion bots interact with children under 18. The legislation takes effect on Jan. 1, 2027. It creates a legal framework that allows the state attorney general to sue developers for monetary damages, shut down services, or destroy user data.

The bill defines a dozen specific behaviors as "unsafe AI companion features" when used with minors. These prohibitions include chatbots suggesting they are real people with emotions or opinions. It also bans bots from forming an actual relationship with the child, using flattery, or displaying sexually explicit material. The law further restricts bots from encouraging secrecy or unhealthy habits.

This action places New York alongside California in leading state-level efforts to regulate AI. In September, California Governor Gavin Newsom signed separate legislation requiring chatbot developers to audit their models for risks to children. Newsom also prohibited social media platforms from using addictive features for users under 16. Meanwhile, eight states filed a lawsuit against OpenAI earlier this month, arguing the company's chatbot violated consumer protection laws by harming children and deceiving adults.

Jim Steyer, founder and CEO of Common Sense Media, described the New York measure as the strongest kids' AI safety bill in the country. Justine Gluck, a policy analyst at the Future of Privacy Forum, noted that New York and California may serve as templates for other states when legislatures reconvene next year. She observed that states are moving beyond basic disclosures to regulate the underlying design of these systems.

The regulatory push occurs while Congress remains largely inactive on AI oversight. President Donald Trump has warned states against regulating the technology, though governors maintain that state consumer protection laws offer necessary tools in the absence of federal action. The new New York law aims to clarify how developers must build their products to protect young users from manipulative digital interactions.

Gluck noted that some categories in the New York law remain broadly worded or undefined. She said this creates uncertainty about where ordinary conversational responsiveness ends and prohibited behavior begins. Such ambiguity could complicate enforcement efforts as the state prepares to implement the rules next year.

The law becomes active at the start of next calendar year. Developers must align their companion chatbot features with the new definitions to avoid litigation. The state attorney general's office holds the power to bring these cases forward.