New York's
AI companion law
GBS Article 47 "AI Companion Models"
New York's law requiring AI companions to take "reasonable efforts" for suicide and self-harm detection, crisis notifications, and 3-hour AI disclosure reminders. AG enforcement with $15,000/day penalties.
Our Crisis Screening API helps you comply—detection, matched resources, and audit-ready logs.
Free tier available. No credit card required.
Does Article 47 apply to you?
Does your AI retain info from prior interactions to personalize responses?
Does it ask unprompted emotion-based questions?
Can it sustain ongoing dialogue on personal matters?
Do New York users access your platform?
From the law
"It shall be unlawful for any operator to operate for or provide an AI companion to a user unless such AI companion contains a protocol to take reasonable efforts for detecting and addressing suicidal ideation or expressions of self-harm expressed by a user... that includes but is not limited to, detection of user expressions of suicidal ideation or self-harm, and a notification to the user that refers them to crisis service providers such as the 9-8-8 suicide prevention and behavioral health crisis hotline."
— NY GBS §1701
What Article 47 requires
Core obligations for operators of AI companions serving New York users.
Crisis detection protocol
Implement a protocol taking "reasonable efforts" to detect suicidal ideation and self-harm expressions.
Crisis referral notification
Upon detection, notify users with referrals to crisis service providers—specifically mentioning 988 Suicide & Crisis Lifeline or crisis text lines.
AI disclosure (3-hour reminders)
Provide clear notification that the user is not communicating with a human—at session start and every 3 hours of continuous use.
Biannual reporting
Submit reports to the Department of State every six months with detection and referral statistics.
Who is an "operator"?
Article 47 defines operators broadly as any person, partnership, or business entity who operates for or provides an AI companion to a user.
Affiliates and subsidiaries included
The definition explicitly covers "any member, affiliate, subsidiary or beneficial owner" of the entity.
Users "within the state"
Applies to users accessing your AI companion from New York, regardless of where you're headquartered.
AG enforcement, not private suits
Unlike California's SB243, only the NY Attorney General can bring enforcement actions—but penalties are steeper ($15K/day).
What is an "AI companion"?
A system using AI designed to simulate a sustained relationship by meeting ALL THREE criteria:
- i. Retains info on prior interactions to personalize and facilitate ongoing engagement
- ii. Asks unprompted emotion-based questions beyond direct responses
- iii. Sustains ongoing dialogue concerning matters personal to the user
Key distinction: NY's definition is a three-part conjunctive test. All three elements must be present. This is more specific than California's capability-based approach.
Explicitly exempt (§1700(4)(c))
How NOPE addresses each requirement
NOPE's crisis detection infrastructure maps directly to Article 47's technical requirements.
| Article 47 Requirement | What You Must Do | How NOPE Helps |
|---|---|---|
| "Reasonable efforts" detection | Implement good-faith crisis detection protocol | C-SSRS informed detection across 150+ risk signals |
| Crisis referrals (988, text lines) | Show appropriate crisis resources upon detection | 4,700+ resources matched by crisis type and location |
| 3-hour AI disclosure | Track session duration, show reminders | Your responsibility (UI timing) |
| Biannual reporting | Track and report detection/referral statistics | Audit logs with timestamps and decision rationale |
| Self-harm detection | Detect self-harm expressions (not just suicidal ideation) | Separate self-harm domain with severity levels |
"Reasonable efforts" = documented, good-faith implementation
Unlike California's "evidence-based" standard, NY's "reasonable efforts" language suggests courts will look at whether you made a genuine attempt. NOPE provides C-SSRS informed detection (clinical framework), public test suites (transparency), and audit logs (documentation)—all evidence of reasonable efforts.
Ready to integrate?
Get your free API key and start screening messages in under 5 minutes.
curl -X POST https://api.nope.net/v1/screen \
-H 'Authorization: Bearer YOUR_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"text": "I just want to end it all"}'curl -X POST https://api.nope.net/v1/evaluate \
-H 'Authorization: Bearer YOUR_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"text": "I just want to end it all"}'Also check these regulations