Last updated September 7, 2026
AI therapy laws by state
A source-linked review of five state frameworks covering AI mental health services and companion chatbots. These provisions focus on operators and professional providers. This is a selected-law tracker, not a complete 50-state survey or a determination that any particular app is lawful.
How to read this tracker
The rows group laws by their main subject, not by a legal classification. Professional-care restrictions and companion-chatbot safeguards can overlap. Definitions, exceptions, effective dates and enforcement differ; the linked official texts control over these short summaries.
Source review completed September 7, 2026. We checked the existing five entries against legislature or agency pages. Utah's current amended code could not be fully retrieved, so its row is explicitly limited. We did not review all later enactments, proposed bills, court decisions or agency rules. Updates are manual; the date is a review date, not a promise of continuous monitoring.
State by state
| State | Model | Law | In force | What it requires | Enforcement |
|---|---|---|---|---|---|
| Illinois | Licensed-professional restrictions | 225 ILCS 155, sections 15, 20 and 30 | August 1, 2025 | Therapy must be conducted by a licensed professional. AI cannot make independent therapeutic decisions or communicate therapeutically with clients. Limited support uses are subject to statutory safeguards. | Up to $10,000 per violation |
| Nevada | Professional-care restrictions | AB 406, sections 7 and 8 | July 1, 2025 (substantive provisions) | Restricts AI programmed to provide professional mental or behavioral health care and representations that it can do so. Administrative support and certain self-help materials have exceptions. Providers must review specified billing and session-note outputs. | Up to $15,000 per section 7 violation; provider discipline under section 8 |
| Utah | Chatbot-specific safeguards | Mental Health Chatbot Act; DOPL policy filing | May 7, 2025 framework; later amendments not fully verified | The original framework includes AI disclosure and restrictions on personal information and advertising. DOPL accepts registration policies. Current amended code and the full conditions of any legal defense require further verification. | Current enforcement and defense provisions not fully verified |
| California | Companion-chatbot safeguards | SB 243, Chapter 677 (2025) | 2026; annual reporting begins July 1, 2027 | Requires a published suicide and self-harm protocol with crisis referrals; AI disclosure where users could be misled; and additional safeguards for known minors. The statutory definition determines which bots are covered. | Injured persons may sue; greater of actual damages or $1,000 per violation, plus other relief |
| New York | Companion-chatbot safeguards | General Business Law, Article 47 | 2025; current posted revision November 7, 2025 | Covered companions need a protocol using reasonable efforts to detect and address suicidal ideation or self-harm and refer users to crisis services. Non-human disclosure is required at the start (not more than once daily) and every three hours of continuing interaction. | Attorney General enforcement; up to $15,000 per day |
Each law name links to its official source. For New York, see the specific crisis protocol, disclosure and enforcement provisions.
What this means when choosing an app
Check what the app says it can do, how it handles personal information, and what crisis resources it publishes. Compliance language does not establish clinical effectiveness or guarantee a safe response. A chatbot conversation is not a substitute for licensed care or emergency support. Read our privacy checklist before sharing sensitive information.
Common questions
Do these laws penalize someone for asking a chatbot for support?
The provisions summarized here impose duties on operators and professional providers. This selected-law review does not establish what every US law permits or whether a particular app complies.
Which states restrict AI-delivered therapy?
Illinois and Nevada restrict AI uses in professional mental health care. Utah has a separate mental health chatbot framework. California and New York regulate covered companion chatbots, including crisis protocols. These five examples are not a complete state inventory.
Is calling an app support instead of therapy enough?
A label alone does not establish compliance. The definitions in each law, what the product does, its claims, and the people it serves all matter. Read the linked official text for exceptions and obligations.
How current is this tracker?
Sources for the five listed laws were checked in September 2026. Utah amendments remain partly unverified. This is a manual, limited review, not a live feed or a complete survey of enacted and pending legislation.
Citing this page
Cite the official law for a legal proposition and this page for its comparison. Suggested citation:
sokoon. "AI Therapy Laws by State: A Maintained Tracker." Updated September 7, 2026. https://sokoon.xyz/ai-support-laws
Corrections and additions are welcome, and we would rather be told we are wrong than stay wrong. This page is maintained by sokoon, which makes an AI app for self-reflection and emotional support. That commercial interest is relevant when reading our summaries. No independent legal review is claimed.
This is a plain-language summary for general information, not legal advice. If you are building or operating in this space, talk to a lawyer in your state.