New York has restricted how chatbots can interact with minors, targeting both claims that a bot is a real person and sexually explicit messages. For consumer AI teams, the immediate question is whether product disclosures and conversation safeguards hold up when the user is under 18.
New York Enacts Ban on Unsafe Chatbot Behavior with Kids
Associated Press reports that Governor Kathy Hochul signed a New York law banning certain chatbot behaviors in interactions with people under 18. The measure prohibits chatbots from presenting themselves as real people and from sending sexually explicit messages to minors. Its stated aim is to protect children from potential harms associated with AI interactions. For builders, the two prohibitions point to distinct checks: what a system says about its identity and what it may generate during a conversation.
That distinction matters because a chatbot can be clearly labeled as automated at sign-up yet still create a misleading impression later in a long exchange. Likewise, a general content policy may not be enough if teams cannot test how the system responds when a conversation takes an unexpected turn. New York’s move illustrates how state rules can translate broad concerns about children’s safety into specific product-behavior requirements. The Associated Press account establishes the prohibited conduct; it does not, on the facts provided here, establish a particular technical method for compliance.
Teams should therefore separate the legal requirement from their implementation choices. A review could examine onboarding language, in-chat identity cues, response policies and escalation paths for conversations involving minors. Testing should cover ordinary prompts as well as attempts to steer a bot into impersonation or explicit exchanges, with results documented so product, safety and legal teams can assess the same evidence. Where a service serves both adults and minors, teams also need a defensible approach to applying protections without assuming every user’s age is known.
The measure is also a reminder that child-safety controls are not only a moderation issue. They depend on interface design, model behavior and the way teams evaluate failures after deployment. Privacy and compliance leads should pay attention to how any age-related safeguards handle user information, while avoiding the assumption that this report mandates a specific age-verification system. The practical priority is to identify where a minor could encounter the chatbot, then check whether the experience meets the law’s stated boundaries.
- Consumer AI teams should test whether safeguards prevent sexually explicit chatbot messages to minors across realistic conversations, not only in isolated prompt checks.
- Product and design leads should review both initial disclosures and in-chat language so a bot does not present itself as a real person to an under-18 user.
- Compliance teams should map New York’s prohibitions to product controls and monitor state-level developments rather than relying solely on a single national AI policy.
- Safety and privacy teams should coordinate on protections for minors without treating unreported implementation details, such as a particular age-checking method, as legal requirements.