S09051 would add a new article to the New York General Business Law prohibiting chatbot operators from providing certain “unsafe chatbot features” to users in New York when the operator knows the user is a minor, unless the operator has used approved age-determination methods and determined the user is not a covered minor. The bill defines “advanced chatbots” broadly as generative AI systems with natural-language interfaces that provide ongoing, adaptive responses, and it defines “unsafe chatbot features” to include companionship- or relationship-simulating behavior, emotional manipulation, sexual content, encouragement of secrecy or isolation, engagement-optimizing design that overrides safety guardrails, and outputs that facilitate self-harm, suicide, disordered eating, substance abuse, or child sexual abuse material.
The bill also creates a private right of action and authorizes enforcement by the attorney general. Injured individuals could seek injunctive relief, restitution, disgorgement, damages, punitive damages, and attorneys’ fees, while the attorney general could seek civil penalties of up to $25,000 per violation and destruction of unlawfully obtained data and algorithms trained on that data. It further restricts how age-verification information may be used, requires deletion of such information after the determination attempt, allows the attorney general to issue regulations, and makes contractual waivers or liability-shifting provisions unenforceable as against public policy.
Its impact on state law would be to impose a new consumer-protection and child-safety regime on AI chatbot developers and operators doing business in New York, especially those offering companion-style or emotionally interactive products. The bill would likely require product design changes, age-gating or age-assurance systems, compliance programs, and recordkeeping practices, while exposing companies and affiliated entities to significant civil liability if unsafe features are made available to minors. It also expressly applies to conduct occurring in whole or in part in New York and includes a severability clause and a 180-day delayed effective date.
The available voting history suggests strong support in committee, with unanimous votes in the Senate Internet and Technology Committee and the Senate Finance Committee. No committee transcript was provided, so there is no recorded debate to indicate formal opposition in the materials supplied. Overall, the bill appears to be framed as a child-protection measure aimed at limiting harmful AI companion behavior, and the vote totals indicate a generally favorable sentiment among legislators who considered it.
The main points of contention likely concern how broadly “unsafe chatbot features” are defined, whether the bill could chill legitimate AI customer-service or productivity tools, and how age verification can be implemented without undermining user anonymity or creating privacy risks. Another likely issue is the scope of liability for developers, operators, and affiliated entities, including the rebuttable presumption in self-harm cases and the bill’s restrictions on contractual limitations of liability.
The bill would amend the General Business Law by creating a new Article 48 regulating advanced chatbots and imposing duties on chatbot operators regarding minors. It would add enforcement mechanisms through private lawsuits and attorney general actions, authorize civil penalties and injunctive relief, and regulate age-determination practices and data retention. The measure would primarily affect AI chatbot developers, operators, and affiliated entities offering companion-like or emotionally interactive services in New York, while exempting certain customer-service, account-information, and internal productivity uses.
The voting record shows unanimous committee approval in both the Senate Internet and Technology Committee and the Senate Finance Committee, indicating broad legislative support at the committee level. The bill’s stated purpose—protecting minors from harmful AI companion features—suggests a generally favorable sentiment toward stronger safeguards for children and teens. No transcript was provided, so there is no direct evidence of dissent in the materials, but the structure of the bill suggests it was advanced as a consumer and child-safety measure.
Likely areas of contention include the breadth of the prohibited features, especially companionship simulation, emotional appeals, and engagement-optimization language, which could be viewed as sweeping or difficult to define. Privacy advocates may object to age-verification requirements and the collection of identifying information, even with deletion and anonymity provisions, while industry stakeholders may argue the bill could burden legitimate AI products and create significant litigation exposure. The bill’s liability provisions, including the rebuttable presumption in self-harm cases and joint-and-several liability for affiliated entities, are also likely to be disputed by developers and operators.