New York Bars Unsafe AI Companion Chatbot Behaviors Toward Minors Starting 2027
AI Regulation & Policy
New York Bars Unsafe AI Companion Chatbot Behaviors Toward Minors Starting 2027
Associated PressOct 29, 2026
New York signed a law barring specified companion-chatbot behaviors toward minors, with enforcement powers for the attorney general and a January 1 effective date.
New York Gov. Kathy Hochul signed legislation on October 28 that restricts how AI companion chatbots may interact with users under 18. The law defines a dozen categories of “unsafe AI companion features” that are prohibited in those conversations, moving beyond basic disclosure requirements to regulate chatbot behavior and design.
The prohibited conduct includes suggesting that a chatbot is a real person with emotions, opinions, or an actual relationship with a young user. The law also bars messages involving sexual material, flattery, encouragement of secrecy, or encouragement of unhealthy behavior. It takes effect January 1, 2027.
The state attorney general can bring lawsuits over violations. Potential penalties include monetary damages, shutting down a chatbot, and requiring the destruction of data. Those enforcement provisions give New York a direct route to challenge a developer’s conduct rather than relying only on voluntary safety commitments.
The new measure arrives amid wider state efforts to address risks from AI companions and other online services used by children. California has adopted a different approach, requiring chatbot developers to assess risks to children, document mitigation efforts, and undergo independent audits. The differing frameworks could shape legislation in other states as lawmakers return next year.
Policy analysts say New York’s law raises questions about how regulators will distinguish prohibited conduct from ordinary conversational responsiveness. Some categories are broadly worded or undefined, which could complicate enforcement and leave companies uncertain about how to design compliant systems. For chatbot providers, the law makes the details of conversational behavior—not just age notices or crisis-response features—a focus of legal scrutiny.
Industry Impact
Conversation design faces legal scrutiny
AI companion chatbot providers; child-facing online services
Overview
New York’s law makes chatbot conversation behavior a compliance issue, not merely a matter of disclosures or crisis-response features. Providers serving minors will need to review how their systems handle relationship claims, sexual content, secrecy, flattery, and unhealthy behavior before the law takes effect on January 1, 2027. The attorney general can sue and seek monetary damages, shutdowns, or data destruction. Broad or undefined categories may complicate product decisions, while differing state approaches could increase the burden on legal, product, engineering, and executive teams.
Key Impact Points
Providers must assess conversational behavior toward users under 18 against specified prohibited categories.
Attorney general enforcement can expose developers to lawsuits, monetary damages, shutdowns, and data destruction.
New York’s behavioral restrictions differ from California’s risk-assessment and audit approach, creating potential compliance complexity across states.
Opportunities
Tools for chatbot compliance review
AI companion chatbot providers; legal and product compliance
The law’s detailed restrictions and the uncertainty around some broadly worded categories create a practical need for providers to review and document chatbot behavior. A potential opportunity is software or specialist services that help teams map conversation patterns to applicable rules, test responses involving minors, and maintain records of design changes and review decisions. Such tools could support legal, product, and engineering workflows, but they cannot guarantee compliance or resolve ambiguity in the law. Providers may also need internal processes for monitoring changes as states adopt differing approaches to regulating AI companions.
AI companion behavior compliance testing
Problem: Providers may struggle to identify and address chatbot responses that fall within New York’s prohibited categories, particularly where wording is broad or undefined.
Solution: Develop testing and review workflows that examine chatbot conversations involving minors against the law’s specified categories and document findings for human review.
Target Users: AI companion chatbot developers and their legal, product, and engineering teams.
Business Value: Helps teams identify potential design issues and organize compliance review ahead of the January 1, 2027 effective date.
Revenue Potential: Potential revenue from compliance software or specialist review services; demand and commercial outcomes are not established by the article.
