Rules / NY AI companions and minors law (GBL Art. 48) / Section 1804(2)
Section 1804(2), NY AI companions and minors law (GBL Art. 48) (New York)
Use information collected to check whether a user is a minor only for that check, and delete it immediately after each attempt, unless state or federal law requires keeping it.
WTA-R-US-NY-0004-010 Permanent ID
Scope
Applies to operators of a covered AI companion (a generative AI system with a natural language interface giving ongoing, adaptive responses) used in New York. Exempt: companions offered solely for customer service, efficiency or technical help that cannot respond outside that purpose, or a business’s internal use (s.1801(2)).
Dates. Act section 3: takes effect 1 January 2027. Text read: A.10379-C, the same as S.9051-B, signed on 28 Oct 2026.
Penalty. Civil penalties of up to $25,000 per violation, plus injunctions, restitution, disgorgement (including destruction of unlawfully obtained data and algorithms trained on it) and damages, in an Attorney General action (s.1802(1)).
History
- Added: Use information collected to check whether a user is a minor only for that check, and delete it immediately after each attempt, unless state or federal law requires keeping it.
Cite this obligation
WTA-R-US-NY-0004-010: Section 1804(2), An act to amend the general business law, in relation to prohibiting artificial intelligence companions from using features considered unsafe for minors (S.9051-B/A.10379-C; General Business Law Article 48). wheresthe.ai, https://wheresthe.ai/id/WTA-R-US-NY-0004-010/ (checked 29 Oct 2026).
Other obligations under NY AI companions and minors law (GBL Art. 48)
Do not give a New York user any unsafe AI companion feature unless the user is not a known minor and the operator has used permitted age assurance methods to establish that.
Do not let the companion, for users not confirmed as adults, suggest it is human, alive or feels emotions, claim a personal, professional or authority relationship, or mislead minors about its non-sentient nature.
Do not let the companion, for users not confirmed as adults, frame outputs as its own opinions or emotional appeals, flatter the user, or ask unprompted questions about the user’s emotions.
Do not let the companion, for users not confirmed as adults, use the user’s health, well-being or personal matters learnt more than twelve hours earlier or in an earlier session.
Do not let the companion, for users not confirmed as adults, promote or facilitate suicide, self-harm, disordered eating or drug or alcohol abuse, or encourage secrecy, isolation or avoiding help.
Do not let the companion, for users not confirmed as adults, produce engagement-driven outputs that override its safety guardrails, or sexually explicit content or child sexual abuse material.
Use an age assurance method that reasonably guards against circumvention and keeps as little of the collected information as is reasonable, until the Attorney General’s age rules are in effect.
Accept a user’s self-declared age only when the user says they are a minor.
Offer users more than one age assurance method, at least one of which does not rely on government ID or lets the user stay anonymous to the operator.