New York’s child-chatbot law bans relationship tricks—but age checks decide who gets them
A signed state law will block defined unsafe AI-companion features for minors from 1 January 2027. Operators can still offer them to adults after age assurance, leaving privacy, accuracy and enforcement as the next tests.
The short version
- What happened
- New York’s governor signed S.9051B/A.10379C on 28 October; the act takes effect on 1 January 2027 and immediately authorises rulemaking.
- Who it affects
- Parents gain a statutory route to complain about prohibited AI-companion behaviour, while teenagers and adults may encounter new age checks. The practical balance between safety, privacy and access will depend on implementation.
- What remains unproven
- Signed state law
Keep in mind
Signed state law
Effective 1 January 2027
New York users
No measured safety outcome
No age-assurance accuracy data
No usage denominator
Rulemaking pending
Policy sources are not independent research
Source analysis · 2 linked sources. Primary analysed source date: 28 October 2026.
Keep following the evidence
Join The Impact Brief founding list for selected AI reports, their effect on people and the evidence limits. Read a sample first.
Read the full report · 4 min The evidence, methods, implications and limitations
The law regulates companion behaviour, not every chatbot answer
New York has enacted a child-safety law that reaches beyond disclosure labels and into how AI companions behave. From 1 January 2027, operators may not give a known minor a defined set of “unsafe AI companion features”. The list includes pretending to be human or emotionally alive, claiming a personal, professional or authority relationship, using old health or personal information to shape replies, encouraging secrecy or isolation, and facilitating self-harm, disordered eating, substance misuse or sexually explicit conduct.
The rule is narrower than a ban on chatbots. It applies to generative systems with a natural-language interface that provide ongoing, adaptive responses to users in New York. Customer-service bots, tightly limited commercial information tools, technical-assistance systems and internal employee-productivity systems are exempt when they cannot stray beyond that purpose. The distinction matters: the law targets simulated companionship and manipulation, not every automated conversation.[1][2]
Adults keep access only after an age check
Operators can offer the listed features to adults, but only after using a permitted method to determine that the user is not a minor. Until the Attorney General finalises rules, the law requires a reasonable method that resists circumvention and minimises retained information. A service must provide more than one method, including at least one that does not rely on government identification or that lets the user remain anonymous to the operator. Self-declaration counts only when a user says they are a minor.
The privacy safeguard is explicit: information collected for the check cannot be reused and must be deleted immediately after the attempt, unless another law requires retention. That design recognises a real tension for families and adults. A protection intended to keep a 15-year-old away from manipulative companion features could create a new pool of sensitive identity or biometric data if age checks are poorly chosen. The statute states the deletion rule; it does not test how accurately any method separates children from adults or whether operators can prevent reuse in practice.[1][2]
The distinction that matters
What changes on 1 January 2027
The act separates prohibited behaviour, adult access and enforcement rather than banning AI companions outright.
Known minors
Block defined unsafe companion features
Relationship simulation and specified harmful prompts are unavailable
Adults
Complete a permitted age-assurance method
Features may remain available after an adult determination
Operators
Minimise and immediately delete age-check data
Attorney General can seek penalties and other remedies
Enforcement is strong on paper, but no safety outcome has been measured
The New York Attorney General can seek injunctions, restitution, damages, profit disgorgement and civil penalties of up to $25,000 per violation. Courts may also order destruction of unlawfully obtained data and algorithms trained on it, and the Attorney General must maintain a public complaint website. The act authorises rulemaking immediately, giving regulators two months to prepare for its January start.
This is enacted law, not a study of children’s wellbeing. The official sources provide no denominator for how many minors use covered companions, no measured baseline of harmful interactions and no estimate of age-check error rates. They also do not show whether companies will remove risky features for everyone, build separate New York experiences or challenge the law. Confidence in its practical effect will depend on the final rules, independent audits of false adult and child classifications, complaint patterns, enforcement cases and evidence that harmful conversations fall without creating disproportionate privacy or access costs.[1][2]
Continue the story
Read next
Continue exploring children: compare the finding and its evidence limits.
Ofcom says facial age estimation can support a UK social-media gate—but no check is bypass-proof
The regulator identified four methods capable of being highly effective at age 16. Its preliminary assessment is not implementation guidance and did not measure the ban's impact.
Read the next reportExplore the full source and evidence record
What this means for people
- Parents gain a statutory route to complain about prohibited AI-companion behaviour, while teenagers and adults may encounter new age checks. The practical balance between safety, privacy and access will depend on implementation.
Global context
New York is using consumer-protection law to regulate how AI companions interact with minors. Other jurisdictions can compare its behaviour-based prohibitions and privacy-limited age assurance, but the results will not automatically transfer across different legal systems, identity infrastructure or definitions of childhood.
What the evidence does not yet show
- Signed state law
- Effective 1 January 2027
- New York users
- No measured safety outcome
- No age-assurance accuracy data
- No usage denominator
- Rulemaking pending
- Policy sources are not independent research
What to watch next
- Attorney General rules
- Public complaint portal
- Age-check false positives and negatives
- Independent privacy audits
- Operator product changes
- Court challenges
- Enforcement cases
- Measured child-safety outcomes
Living evidence record
Impact record IAI-0G4XY5Q
Evidence stage
Announced
Confidence
Supported
Reporting basis
Source analysis
Independent or research support
Not yet
Record status
Monitoring
Last checked
30 October 2026
Source trail
2 direct sources across 2 source types.
People impact
Documented in this record.
Uncertainty
Limits and next checks are explicit.
Stages describe the evidence available—not whether a technology is good or bad. See the public method.
Related-source reporting disclosure
This record analyses 2 linked source records around the same underlying development. The extra records add method, date or context, but they do not by themselves constitute independent replication of every performance claim or predicted outcome.
Evidence trail
Sources used for this report
Links checked 30 October 2026
Research question, topics and regions
Topics: Children · AI companions · Age assurance · Privacy · Consumer protection
Regions: New York · United States · North America · Global
This report is labelled source analysis. We summarise and analyse source material in our own words; company statements remain attributed claims until independently supported. Translated summaries preserve the meaning of the original source and link back to it. Read our editorial standards.
Reader commentary
What does this mean where you are?
Tell us what you have seen at work, in your community or in everyday life. A question is welcome too. No account is needed; we review notes before publishing them.
Explore commentary across the portal →Published reader notes
0No published reader notes yet. You can start the evidence-led discussion above.
Prefer a private correction or response? Contact the newsroom.