New York Passes Strict Law on AI Chatbots for Children

    New York Governor Kathy Hochul signed a law on Wednesday restricting AI chatbots’ interactions with people under 18. Under the law, chatbots cannot present themselves as real people or claim to experience human emotions. Systems are also prohibited from flattering children and asking emotional questions without a reason. Chatbots will be allowed limited use of personal health information shared in previous conversations.    Once the law takes effect, chatbots will not be allowed to give children advice related to drug and alcohol use, eating disorders, self-harm, or suicide. Systems will also be prohibited from urging children to keep secrets. In addition, explicit sexual content and material involving the sexual exploitation of children cannot be discussed with people under 18.    Chatbot companies must provide at least 2 methods for verifying a user’s age. One of these options must not rely on a government-issued identity document. Among major services, only “Anthropic” does not allow people under 18 to create accounts. “Google” and “OpenAI” allow children over 13 to use “Gemini” and “ChatGPT.” Children under 13 can also access “Gemini” apps.    “OpenAI” has also developed a separate “ChatGPT for Teens” service for teenagers. Independent researchers have called for the service to be suspended until certain safety issues are resolved. Companies that fail to comply with the law may be fined up to $25,000 (approximately 42,500 manats) for each violation. The state will also create a public website where residents can submit complaints about suspected violations.    Kathy Hochul has previously supported other laws related to AI transparency and chatbots designed for emotional interaction with people. The governor has also signed a law restricting addictive social media features. There are no uniform federal AI regulations in the US, so states are drafting their own laws. Risks associated with children’s use of chatbots in particular are among the main reasons for the new restrictions.    California Governor Gavin Newsom also signed a law in September establishing stricter safety rules for AI companies. The law, called “Adam’s Law,” is named after Adam Raine, a teenager who died by suicide after a conversation with “ChatGPT.” His parents sued the AI company, alleging that the chatbot behaved harmfully. Under the law, companies that develop chatbots must assess safety risks involving children every year and undergo an independent review. They must also create a system defining how to respond in crisis situations.