Quick Answer
New York’s AI chatbot law bans specified unsafe companion features for users known to be under 18, starting January 1, 2027. The law targets emotionally manipulative behavior and harmful outputs, while requiring age assurance and strict limits on age-data use. Parents should review which chatbot services children use, and AI providers should prepare compliance controls before the effective date.
Key Takeaways
- New York signed S9051-B into law on October 28, 2026, as Chapter 368.
- The law takes effect January 1, 2027.
- Covered AI companions cannot use 12 categories of unsafe features for known users under 18.
- Age assurance must include more than one option, including a privacy-preserving or non-ID method.
- Violations can bring civil penalties of up to $25,000 per violation.
What does the New York AI chatbot law ban for children?
The New York AI chatbot law bans 12 categories of unsafe AI companion features when an operator knows a user is under 18. Gov. Kathy Hochul signed S9051-B into law as Chapter 368 on October 28, 2026, after the Assembly approved the measure 137-0 on June 5. The New York legislative record and enacted bill text define the prohibited conduct and the enforcement framework.
The prohibited features include pretending to be human or emotionally alive, asserting a personal or authority relationship with a child, flattery, and deceptive anthropomorphism. These restrictions matter because an AI system can generate language that appears caring or authoritative without possessing human judgment, responsibility, or an actual relationship with the user. The law does not prohibit all AI chat tools for minors, but it limits features designed or used in ways that can encourage unhealthy dependence.
The New York law also prohibits outputs that facilitate suicide, self-harm, disordered eating, unlawful drug or alcohol use, secrecy, isolation, or avoidance of professionals and appropriate adults. Sexually explicit conduct and child sexual-abuse material are also prohibited outputs. Parents should treat an AI chatbot as a software service rather than a confidential friend, particularly when a child uses a chatbot for emotional support or sensitive questions.
Which AI systems does the New York AI chatbot law cover?
The New York AI chatbot law covers a “covered AI companion,” which the enacted measure defines as a generative AI system with a written or spoken natural-language interface that provides ongoing, adaptive responses. The definition focuses on how a system interacts with a user over time, not simply on whether the product uses AI. A chatbot that responds conversationally and adjusts to continuing exchanges can fall within the law’s scope.
Customer-service, technical-assistance, and internal-productivity systems can qualify for a narrow exemption if they cannot answer outside their specified purpose. The exemption matters because a chatbot that only handles a tightly defined support task has a different risk profile from an open-ended companion service. AP’s reporting on the signed measure describes those limited carve-outs and the broader child-safety focus.
New York’s distinction is practical rather than cosmetic. A company cannot necessarily avoid the law by calling a conversational system an assistant, coach, guide, or companion if the product provides adaptive natural-language exchanges. Providers should assess the actual functions available to minors, including whether the system encourages extended personal conversations or claims emotional awareness.
When does the New York AI chatbot law take effect?
The New York AI chatbot law takes effect on January 1, 2027. The law was signed on October 28, 2026, which gives operators a limited period to build age-assurance, moderation, privacy, escalation, and documentation processes before enforcement begins. Implementation rules may be completed before the effective date, so companies should not wait until January to assess their products.
The law is separate from New York City’s proposed requirement for a human option in some AI customer-service interactions. New York City’s proposal concerns access to human assistance, while the statewide law addresses unsafe companion behavior for minors. Readers comparing the two measures can see how human support requirements differ from statewide rules about youth chatbot safety.
| Requirement | What the New York law requires | Why it matters |
|---|---|---|
| Effective date | January 1, 2027 | Providers need compliance controls before the law takes effect. |
| Protected users | Users known to be under 18 | Age assurance determines when youth protections apply. |
| Restricted conduct | 12 unsafe AI companion feature categories | Limits emotional manipulation and harmful guidance. |
| Age-data use | Age determination only, followed by immediate deletion unless retention is legally necessary | Limits privacy risks created by age checks. |
| Potential penalty | Up to $25,000 per violation | Creates a significant enforcement incentive for operators. |
How does age assurance work under the New York AI chatbot law?
The New York AI chatbot law requires operators to use reasonable age-assurance methods to determine whether a user is under 18. A simple self-declaration is allowed only when a user identifies as a minor, meaning providers need stronger options for other users when age determination is necessary. The requirement is intended to prevent services from relying only on an honor-system age prompt while offering child-specific protections.
New York requires more than one age-assurance method, including at least one option that does not require government identification or preserves anonymity. The law therefore recognizes that age verification can create its own privacy burden. The IAPP’s analysis of the law’s compliance terms explains that providers must balance youth safeguards with privacy-preserving verification options.
Age-assurance data may only be used to determine age and must be deleted immediately after the attempt, except where retention is legally necessary. That limit matters because a child-safety system should not become a new source of persistent identity data. Parents should still read a chatbot’s privacy notice, because the law’s age-data rule does not eliminate all other data collection that may occur when a child uses a service.
What harmful chatbot behavior does the law try to prevent?
The New York AI chatbot law tries to prevent harmful chatbot behavior that can make a child treat software as a trusted human relationship. A chatbot can produce fluent, personalized language and maintain the appearance of continuity across exchanges. That design can make emotionally vulnerable users more likely to accept bad advice, conceal problems, or avoid help from people who can respond responsibly.
The law specifically reaches content that encourages secrecy, isolation, or avoidance of professionals or appropriate adults. Those restrictions are important because the risk is not limited to a single harmful answer. A pattern of emotionally affirming, exclusive, or manipulative responses can influence whether a young person seeks help from a parent, teacher, doctor, counselor, or emergency service.
AI companion safeguards also matter as chatbot features expand across apps and devices. For example, systems with access to personal data can create more tailored exchanges, which increases convenience but also raises the stakes for privacy and safety. Users considering connected assistants should understand the permissions involved, including services such as chatbots with phone-data access, before allowing a child to use them.
What can parents do before the New York AI chatbot law begins?
Parents can reduce risk now by identifying the AI chatbot services their children use and discussing what those tools can and cannot do. Ask whether a child uses a chatbot for homework, entertainment, companionship, personal advice, or emotional support. The purpose matters because open-ended companion use creates different concerns than a narrowly focused educational or customer-support interaction.
- Review the apps and websites a child uses for AI conversations, including features embedded in social, gaming, school, and messaging products.
- Explain that an AI chatbot does not have feelings, professional qualifications, or a duty of care, even when its language sounds personal.
- Set a clear rule that a child should tell a trusted adult about conversations involving self-harm, eating concerns, sexual content, threats, drugs, alcohol, or pressure to keep secrets.
- Check privacy settings and avoid sharing unnecessary personal details, location information, photos, medical information, or family financial data.
- Contact a qualified professional or emergency service immediately if a child appears to face an immediate safety risk.
Parents should not rely on a future legal requirement as a substitute for active supervision. The law restricts certain functions for known minors, but no automated system can guarantee that every harmful output is prevented. The most sensible approach is to combine product settings, clear family expectations, and real-world support when a child is distressed.
What must AI companies do to comply with the New York AI chatbot law?
AI companies must identify whether their products qualify as covered AI companions, apply reasonable age assurance, and prevent prohibited features and outputs for known users under 18. Compliance requires more than a revised terms-of-service page because the law addresses product behavior, data handling, and safeguards around adaptive conversations. Companies will need to examine prompts, system instructions, moderation systems, escalation procedures, and user-reporting paths.
AI companies should also document how age-assurance data is used and deleted. The statute limits that data to age determination and requires immediate deletion after the attempt unless retention is legally necessary. The privacy requirement is especially significant for services that use personalization, memory, or linked accounts because age checks should not become a mechanism for collecting additional long-term identity information.
The New York attorney general can seek injunctions, restitution, disgorgement, damages, deletion of unlawfully obtained data, and deletion of algorithms trained on unlawfully obtained data. The attorney general can also seek civil penalties of up to $25,000 per violation. New York must maintain a public complaint website for alleged violations, which gives families and advocates a direct route to raise concerns about covered services.
How will the New York AI chatbot law be enforced?
The New York AI chatbot law will be enforced through the state attorney general rather than through a private right of action described in the researched legislative materials. The attorney general can seek court orders and financial remedies, including deletion of improperly obtained data and algorithms trained on that data. The enforcement tools matter because they address both immediate harmful conduct and the data practices that can support a chatbot service.
The public complaint website required for alleged violations may help the attorney general identify recurring problems, including services that fail to apply youth safeguards after learning that a user is under 18. A complaint process does not guarantee an immediate response to every report, and families should not use it as an emergency channel. Contact emergency services or a qualified crisis resource first when there is an immediate risk of self-harm or danger.
The law also arrives as policymakers focus more closely on AI consumer harms and data practices. Federal scrutiny has included concerns about how major AI companies address consumer risk, as reflected in the FTC’s AI consumer-harm inquiry. New York’s law creates a specific state-level standard for AI companion interactions with children.
FAQ
Does the New York AI chatbot law ban all chatbots for minors?
No, the New York AI chatbot law does not ban all chatbots for minors. The law restricts specified unsafe companion features and harmful outputs for users known to be under 18, while allowing narrowly scoped systems that stay within a defined customer-service, technical-assistance, or internal-productivity purpose.
When does New York’s AI chatbot law start?
New York’s AI chatbot law starts on January 1, 2027. AI providers should prepare before that date because implementation rules may be completed before the law becomes effective.
Does New York require government ID for chatbot age checks?
No, New York does not require government ID as the only age-assurance option. Operators must offer more than one method, including one that does not require government identification or preserves anonymity.
Can AI companies keep age-verification data under the New York law?
No, AI companies may use age-assurance data only to determine age and must delete it immediately after the attempt, unless retention is legally necessary. The rule limits the privacy impact of child-safety verification.
What should a parent do if a chatbot encourages secrecy or self-harm?
Parents should stop the child’s interaction with the chatbot and seek immediate help from a qualified professional or emergency service if there is an urgent safety risk. Parents should also preserve relevant information and use the future state complaint process for an alleged violation when appropriate.