Summary
Chapter 368 adds GBL Article 48, restricting unsafe AI companion features for minors, requiring age assurance, and authorizing Attorney General enforcement.
Healthcare Implications
Consumer AI operators face restrictions on self-harm encouragement, emotional manipulation, and reuse of health disclosures. Narrow customer-service, technical-assistance and internal-business uses are excepted; Article 47 remains intact.
Operational Implications
- From January 1, 2027, operators may provide unsafe features only to users determined to be adults through permitted age assurance. Restrictions address simulated relationships, flattery, self-harm/eating-disorder promotion, isolation, sexual content, and prior-session health/personal information. Pending rules, offer multiple age-assurance methods, including a non-ID or anonymous option; adult self-declaration alone is insufficient. Delete age-check data immediately except where law requires retention. Sole-purpose exceptions apply.