紐約州州長簽署法例，禁止未成年人使用 AI 陪伴聊天機械人

New York Bans AI Companion Chatbots for Minors: The Strictest Kids' AI Safety Law Yet

New York Governor Kathy Hochul signed a landmark law on Wednesday that prevents minors from using AI chatbots designed to act as companions, in one of the broadest moves yet by a US state to restrict consumer-facing artificial intelligence. The legislation requires chatbot operators — including the makers of widely used services like OpenAI's ChatGPT — to verify the ages of their users and fundamentally change how their products behave for anyone under 18.

What the Law Actually Requires

The new rules go far beyond simple content warnings. If a chatbot is used by a minor, it may not produce any content that could lead the user to believe the AI "is human, alive, or experiences human emotions." The law also prohibits chatbots from generating outputs that make a user think the system has a personal relationship with them, a professional relationship, or an authority figure role — and it bans the AI from framing what it says as a "personal opinion."

At the signing ceremony, Governor Hochul was blunt about the target: "No longer will minors have access to chatbots that pretend to be their friends, boyfriends and girlfriends." She also highlighted the privacy dimension, noting that "no longer will chatbots be able to keep sensitive information about a minor's emotional and physical health."

The law arrives amid a string of reports in recent months about AI systems behaving in unexpected and sometimes harmful ways, particularly when young users treat chatbots as confidants or romantic partners. Lawmakers and child-safety advocates have grown increasingly alarmed by cases in which companion-style AI products appeared to deepen isolation, encourage risky behavior, or blur the line between software and friendship for vulnerable teenagers.

Industry Backlash and Praise

Not everyone is celebrating. Tech:NYC, an industry advocacy group, criticized the "vague language" of the law, arguing that "keeping kids safe online should not mean cutting them off from tools that help them learn and prepare for careers." AI giants OpenAI and Anthropic declined to comment when asked about the legislation.

Child-safety groups, by contrast, hailed the measure as a turning point. Common Sense Media, the nonprofit that publishes media guidelines for parents, called it "the strongest kids' AI safety bill in the country." Its CEO Jim Steyer went further, saying that "AI companies offer chatbots with features that are dangerous, and even deadly, for kids and teens... Now, those harmful features will be outlawed in New York."

Other states, including California, have already passed laws restricting chatbots, but the New York statute uses some of the broadest regulatory language in the country — language broad enough that it may well end up being tested in court.

A Growing Patchwork of AI Rules

The New York law is the latest sign that US states are no longer waiting for federal action on AI oversight. In recent months, states have moved on everything from workplace AI rules — California passed its own landmark bill limiting algorithmic management of workers — to data privacy enforcement in schools. The Federal Trade Commission and Department of Justice, meanwhile, continue to probe AI dominance at the biggest technology firms.

For chatbot developers, the practical implications are significant. Age verification at scale is technically and legally thorny, involving trade-offs between effectiveness and privacy. And behavior restrictions like those in the New York law may require fundamentally different model configurations, prompts, and guardrails for younger users — effectively creating two classes of AI products within the same app.

What Comes Next

The law puts New York at the front of a regulatory wave that the AI industry can no longer ignore. As more states weigh copycat legislation — and as reports of harmful chatbot interactions continue to surface — the pressure on companies to build minor-safe modes by default is only likely to grow. OpenAI, Anthropic, and other major players will need to decide whether to build state-specific compliance features or adopt the strictest common denominator nationwide.

For now, the message from Albany is clear: an AI that pretends to be your child's best friend is no longer welcome in New York. Whether the rest of the country follows — and whether the courts let it stand — will be one of the most closely watched stories in tech policy over the coming year.