New York unsafe AI companion features law (2026)
AI-generated text. This page was generated using artificial intelligence.
New York unsafe AI companion features law
New York, United States
Enacted state law
2026 Chapter 368; S9051B; General Business Law Article 48
Signed October 28, 2026, according to the current Senate action record.
Principal provisions January 1, 2027; preparatory rulemaking authorization immediate.
October 28, 2026 — complete active B print and current official signature/actions record.
Overview
New York’s restrictions on unsafe AI companion features were signed as Chapter 368 of the Laws of 2026 on October 28, through S9051B. The principal provisions take effect January 1, 2027; necessary implementing rulemaking is authorized immediately. The signing does not make the principal restrictions immediately operative.[1][2]
Coverage and restricted features
The new Article 48 covers generative-AI systems with a natural-language interface that provide ongoing adaptive responses, including but not limited to AI companions defined in the earlier law. A minor is under 18; a covered minor is a New York user the operator actually knows is a minor. Section 1801 permits the specified unsafe features only if the user is not a covered minor and the operator has used the required age-assurance methods to determine that status.[2]
The feature list includes representations of being human, alive or emotionally sentient; personal, professional or authority relationships; personal opinions or emotional appeals; flattery; unsolicited emotional inquiries; specified reuse of personal or health information from earlier sessions or more than twelve hours earlier; deception; harmful encouragement; secrecy and isolation; engagement that overrides safeguards; and sexually explicit outputs. The attorney general may identify additional deceptive or relationship-simulating features by rule.[2]
The prohibition has specified sole-purpose exceptions for bounded customer service, efficiency/research/technical assistance that cannot answer outside its purpose, and internal business or employee-productivity systems. These exceptions and the broad system definition require attention to the actual service and features.[2]
Age assurance and information safeguards
Pending the specified rules, operators must use reasonable methods that guard against circumvention and minimize retained information. Self-declaration is sufficient only for declaring minor status. More than one method must be available, including at least one that does not rely on government-issued identification or lets the user remain anonymous to the operator.[2]
Age-determination data may be used only for that purpose and must be deleted immediately after the attempt, except as needed to comply with applicable state or federal law or regulation. Article 48 expressly preserves Article 47; it does not replace the earlier companion-law framework.[2]
Enforcement
The attorney general may seek injunctions, restitution, disgorgement, damages, civil penalties up to $25,000 per violation and other relief, including destruction of unlawfully obtained data and algorithms trained on it. The law also requires a public complaint website, voids specified liability-waiver or shifting terms, and allows joint liability for affiliates where the stated deliberate liability-avoidance and frustrated-recovery conditions are established.[2]
The Senate action record also lists approval memo 6. Its separate text and any subsequent chapter-amendment agreement have not been reviewed here; later changes and implementing regulations require separate verification.[1]
Related coverage
References
- ↑ 1.0 1.1 New York Senate, S9051B current status and action history, October 28, 2026 actions: signed Chapter 368, approval memo 6; reviewed October 28.
- ↑ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 S9051B active B print, complete five-page text, §§1–3 and proposed General Business Law §§1800–1805, reviewed October 28, 2026.