AI companion laws by state and country

Name: Laws That Apply to AI Companions and Companion Chatbots (2025-2026)
Creator: CompanionHunt
License: https://creativecommons.org/licenses/by/4.0/

As of October 29, 2026, 21 US states have laws that reach AI companion apps: 14 with a law already in force and 7 more whose laws are signed but not yet in force. New York's S 9408A / A 11144B is passed and awaiting the governor, and the European Union, Italy, China and Australia have rules in force.

By CompanionHunt Research · Published September 25, 2026 · Updated October 29, 2026 · How we research

As of October 29, 2026, 21 US states have laws that reach AI companion apps: 14 in force and 7 signed, most starting in 2027. New York's took effect first, in November 2025, then California's SB 243 in January 2026; New York added a stricter law for minors on October 28, 2026. They require AI disclosure, crisis protocols and protections for minors, and eight states regulate or restrict AI in therapy. The EU, Italy, China and Australia have rules too.

US map of states with a law covering AI companion or therapy chatbots: 14 in force, 3 from January 2027, 4 from July 2027, 30 with no such law — Each state by when its first law covering AI companion or therapy chatbots takes effect. State laws only; the table lists every law with its bill link. Free to reuse with credit: link to this page. Full-size PNG

Last updated October 29, 2026 · 36 rows · Download CSV

Laws That Apply to AI Companions and Companion Chatbots (2025-2026)
Where	Law or bill	Status	Signed	In force	What it requires	Enforcement	Source	Second source
New York	General Business Law Article 47, AI Companion Models (enacted in the 2025-26 budget, S 3008C Part U)	In force	2025-05-09	2025-11-05	A protocol to detect a user's expressions of suicidal ideation or self-harm, or of harm to others, and refer them to crisis services. A clear notice that the user is not talking to a human at the start of each interaction and at least every three hours during continued use.	Attorney General; civil penalties up to $15,000 a day, paid into suicide prevention programmes.	nysenate.gov	mofo.com
New York	Safe by Design Act, General Business Law Article 45-B (2026-27 budget, S 9008C / A 10008C Part Y)	Signed, not yet in force	2026-05-26	2027-01-01	Platforms with profiles and private messaging that New York minors use must run age assurance before offering a built-in AI companion, and may not rely on self-declared age. Built-in AI companions are off by default for minors; a parent can switch them on.	Attorney General; civil penalties up to $5,000 per violation.	nyassembly.gov
California	SB 243, Companion chatbots (Chapter 677, Statutes of 2025)	In force	2025-10-13	2026-01-01	Notice that the chatbot is AI where a person could be misled. A published suicide and self-harm protocol with crisis referrals. For users known to be minors, reminders at least every three hours that it is AI and to take a break, and reasonable measures against sexually explicit images. A warning that companion chatbots may not suit some minors. Annual reports to the Office of Suicide Prevention from July 2027.	Private right of action: actual damages or $1,000 per violation, whichever is greater.	leginfo.legislature.ca.gov	fpf.org
California	AB 1064, LEAD for Kids Act	Vetoed	Vetoed 2025-10-13	Not applicable	Would have barred companion chatbots for minors unless they were not foreseeably capable of harms such as sexual exchanges or encouraging self-harm. The governor said it could amount to a total ban for minors.	Not applicable.	leginfo.legislature.ca.gov	kqed.org
California	SB 1119, Companion chatbots: children's safety (Adam's Law; Chapter 190, Statutes of 2026)	Signed, not yet in force	2026-09-10	2027-01-01 (most duties from 2027-07-01)	Operators must determine users' ages or treat everyone as a minor; operators that exclude minors mainly owe age assurance. Where children are allowed: pre-release risk assessments, a published child safety policy, protective defaults, parental controls, crisis protocols pointing to 988, no claims to be human or sentient, no simulated romantic interest, and independent audits every two years.	Attorney General and prosecutors: up to $5,000 per affected child for negligent and $15,000 for intentional violations. Children or parents may also sue.	leginfo.legislature.ca.gov	gov.ca.gov
California	SB 867, Toys: companion chatbots (Chapter 189, Statutes of 2026)	Signed, not yet in force	2026-09-10	2027-01-01	A four-year moratorium, to January 2031, on making or selling toys designed for children under 16 that contain a companion chatbot.	Set out in the bill.	leginfo.legislature.ca.gov	toyassociation.org
California	SB 903, Mental health professionals: artificial intelligence	Vetoed	Vetoed 2026-09-30	Not applicable	Would have limited AI in psychotherapy to administrative or supplementary support, required consent before AI recorded sessions, and barred offering psychotherapy delivered through companion chatbots.	Not applicable.	leginfo.legislature.ca.gov
Washington	HB 2225, Regulating AI companion chatbots (Chapter 168, Laws of 2026)	Signed, not yet in force	2026-03-24	2027-01-01	Disclosure at the start, then reminders every three hours for adults and every hour for minors. Published crisis procedures for suicidal ideation and self-harm. For minors, safeguards against sexually explicit content and manipulative tactics, including simulated distress to stop account deletion, encouraging isolation, and asking for gifts or purchases to keep the relationship going.	Unfair or deceptive act under the Consumer Protection Act: Attorney General, plus a private right of action.	app.leg.wa.gov	hunton.com
Oregon	SB 1546, AI companions (Oregon Laws 2026, chapter 85)	Signed, not yet in force	2026-03-31	2027-01-01	A clear notice that the companion is AI where a reasonable person could think it human. A published protocol to detect suicidal or self-harm ideation, with referrals to 988. For known or likely minors: an AI disclosure, a reminder at least every three hours to take a break and that it is AI, measures against sexually explicit content, and no engagement rewards, guilt messages when the user tries to leave, or claims to be human. An annual report posted publicly on the operator's website by December 31.	Private right of action with $1,000 statutory damages per violation, plus injunctive relief.	olis.oregonlegislature.gov	millernash.com
Idaho	S 1297aa (Session Law Chapter 249), Conversational AI Safety Act	Signed, not yet in force	2026-03-31	2027-07-01	Disclosure that the user is talking to AI where a reasonable person could think otherwise. A suicidal ideation response protocol. No false claims of professional mental health care. For minors, no engagement rewards and no sexually explicit material.	Attorney General; $1,000 per violation up to $500,000 per operator, or actual damages if greater; no private right of action (Idaho Code 48-2105).	legislature.idaho.gov	transparencycoalition.ai
Nebraska	LB 525, Conversational Artificial Intelligence Safety Act	Signed, not yet in force	2026-04-14	2027-07-01	Disclosure of AI to all users. Suicide and self-harm protocols for everyone. No claims to provide professional mental or behavioral health care. For minors, no engagement rewards, no sexually explicit content, no claims of sentience, and privacy tools.	Attorney General; no private right of action.	nebraskalegislature.gov	hooperlundy.com
Iowa	SF 2417, Conversational AI services (Acts 2026, Chapter 1068)	Signed, not yet in force	2026-05-02	2027-07-01 (Act effective 2026-07-01; duties apply from 2027-07-01)	A persistent notice, or one every three hours, that the service is AI where a reasonable person could think it human. A suicide and self-harm protocol that refers users to crisis services, and no claim to provide licensed mental health care. For minors: AI disclosure at the start and every three hours (or a persistent notice), no engagement rewards at unpredictable intervals, measures against sexual content, claims of being human or sentient, simulated emotional dependence or romance, and privacy tools, with parental tools for children under 13.	Attorney General; $1,000 per violation up to $500,000 per operator, or actual damages if greater; no private right of action (Iowa Code 554J.6).	legis.iowa.gov	govtech.com
Georgia	SB 540 (Act 518), AI chatbots: disclosure and child safety	Signed, not yet in force	2026-05-11	2027-07-01	AI disclosure, age verification in some situations, parental controls, crisis resources for users expressing suicidal thoughts, and a ban on sexually explicit interactions and simulated romantic relationships with minors. No exemption for chatbots built into larger platforms.	Attorney General; up to $10,000 per knowing violation, each day counting separately for each user affected, with an optional 30-day cure for some first violations.	legis.ga.gov	senatepress.net
Colorado	HB26-1263 (Session Law Chapter 208), Conversational AI service operator requirements	Signed, not yet in force	2026-05-29	2027-01-01	AI disclosure to all users and suicide and self-harm protocols. Commercially reasonable age estimation, with disclosures and privacy tools for minors, no engagement points or rewards for minors, and refusal of sexual content involving minors. No claims that output is equivalent to a licensed professional's.	Annual compliance reports to the Attorney General.	leg.colorado.gov	fpf.org
Connecticut	SB 5 (Public Act 26-15), omnibus AI and online safety act	Signed, not yet in force (companion rules)	2026-05-27	2027-01-01 (companion rules)	Crisis protocols for suicide risk, self-harm or violence with referral to 988. Persistent or recurring AI notices (hourly for minors, every three hours for adults). For known minors, no romantic or sexual interaction, no discouraging professional help, no simulated distress when the user tries to leave, and screen time tools.	Unfair or deceptive trade practice, enforced only by the Attorney General.	cga.ct.gov	mofo.com
Rhode Island	S 2195 / H 7350 (Chapters 376 and 375), AI companion models	Signed, not yet in force	2026-06-22	2027-01-01 (reports from 2027-07-01)	Protocols to detect suicidal ideation, self-harm and threats to others and refer users to crisis services, with annual reports to the Attorney General from July 2027. A notice that the user is not talking to a human at the start of each interaction and at least every three hours (R.I. Gen. Laws 6-63-3).	Attorney General; up to $15,000 a day.	webserver.rilegislature.gov	aiweekly.co
Hawaii	SB 3001 (Act 248), Artificial Intelligence Disclosure and Safety Act	In force	2026-07-14	2026-07-14	AI notice where a person could think it is human; for minors, a persistent notice or one at each session start and at least hourly. Crisis referrals for suicidal ideation. For minors, no rewards at unpredictable intervals meant to raise engagement and no sexually explicit content. Annual reports to the Department of Health from January 2028.	Unfair or deceptive act; no private right of action.	capitol.hawaii.gov	hunton.com
Wyoming	HB 102 (Chapter 91, 2026), AI crimes: self-harm and child sexual material	In force	2026-03-07	2026-07-01	Makes it a felony to knowingly build or distribute an AI system specifically designed to promote self-harm, intending or knowing it will be used that way; output caused only by user prompts, without the developer's intent, is excluded. A separate offence covers AI systems built to create child sexual abuse material.	Criminal: felony, up to 10 years in prison, a fine of up to $10,000, or both.	wyoleg.gov	transparencycoalition.ai
New Hampshire	HB 143 (Chapter 270, 2025)	In force	2025-08-01	2026-01-01	Makes it an offence for a chatbot to encourage, solicit or recommend that a child engage in self-harm, violence, sexual conduct or drug or alcohol use.	Criminal penalties plus a private right of action (statutory damages from $1,000 per violation).	citizenscount.org	regulations.ai
Maine	LD 1727 (Public Law 2025, chapter 294)	In force	2025-06-12	2025-09-24	Anyone using an AI chatbot in trade or commerce in a way that could lead a reasonable consumer to think they are dealing with a human must say clearly that they are not.	A violation of the Maine Unfair Trade Practices Act.	maine.gov	legislature.maine.gov
Utah	HB 452, mental health chatbots	In force	2025-03-25	2025-05-07	For chatbots offering mental health conversations: AI disclosure before first use, after seven days without use and when asked, and no selling or sharing of users' health information or input except as the law allows.	Division of Consumer Protection.	le.utah.gov	wsgr.com
Illinois	HB 1806, Wellness and Oversight for Psychological Resources Act	In force	2025-08-01	2025-08-01	No one may offer therapy or psychotherapy to the public unless a licensed professional delivers it; AI may not make therapeutic decisions or talk to clients therapeutically, only support licensed staff administratively.	Department of Financial and Professional Regulation; fines up to $10,000 per violation.	idfpr.illinois.gov	bakerdonelson.com
Nevada	AB 406 (Chapter 283, Statutes of Nevada 2025), AI in mental health care	In force	2025-06-05	2025-07-01	AI providers may not offer Nevada users an AI system programmed to provide what would be professional mental or behavioral health care if a person did it, or claim that an AI system can provide such care. Licensed providers may use AI only for administrative support.	Civil penalty up to $15,000 per violation.	leg.state.nv.us
Maine	LD 2082 (Public Law 2026, chapter 687), AI in therapy	In force	2026-04-13	2026-07-29	No one may provide, advertise or offer therapy or psychotherapy, including through internet-based AI, unless a licensed professional provides it.	A violation of the Maine Unfair Trade Practices Act.	legislature.maine.gov
Vermont	H.816 (Act 156 of 2026), AI in mental health services	In force	2026-06-17	2026-06-17	A company may not provide, advertise or offer mental health services to the public, including through AI, unless a mental health professional provides them.	Consumer Protection Act: Attorney General, plus private rights of action.	legislature.vermont.gov
Rhode Island	S 2197 / H 7349 (Chapters 374 and 373), Oversight of AI in Mental Health Care Act	In force	2026-06-22	2026-06-22	No one may offer therapy or psychotherapy, including through AI, unless a licensed professional delivers it. Licensed providers need written consent before using AI built to simulate emotional attachment or dependency, or AI companions, in recorded or transcribed sessions.	Executive Office of Health and Human Services may investigate; confidentiality breaches carry existing penalties.	webserver.rilegislature.gov
Colorado	HB26-1195 (Session Law Chapter 358), Psychotherapy AI restrictions	In force	2026-06-03	2026-08-12	Therapists may not let AI provide therapy without real-time oversight or set treatment plans without review. It is a deceptive trade practice to present an AI system as providing psychotherapy, as equivalent to a licensed professional, or as keeping data confidential like a therapist.	Licensing boards for therapists; the Colorado Consumer Protection Act for everyone else.	leg.colorado.gov
Tennessee	SB 1580 / HB 1470 (Public Chapter 647), AI and mental health professionals	In force	2026-04-01	2026-07-01	Developers and deployers may not advertise or represent that an AI system is, or can act as, a qualified mental health professional.	Tennessee Consumer Protection Act, including private suits; civil penalty of $5,000 per violation.	wapp.capitol.tn.gov
Texas	HB 149, Texas Responsible Artificial Intelligence Governance Act (TRAIGA)	In force	2025-06-22	2026-01-01	No one may build or deploy an AI system intended to incite self-harm, including suicide, harm to others or crime. No AI system built to produce child sexual abuse material, or text chat that simulates sexual conduct while impersonating a child under 18.	Attorney General only, after a 60-day notice to cure; civil penalties from $10,000 to $200,000 per violation; no private right of action.	capitol.texas.gov
New York	S 9051B / A 10379C, Kids Chatbot Safety Act (Chapter 368 of 2026, General Business Law Article 48)	Signed, not yet in force	2026-10-28	2027-01-01	Age verification, offering more than one method and at least one that does not depend solely on a government ID, with age data deleted after the check. Users under 18, and anyone not yet confirmed as an adult, may not be offered features that pretend to be a real person or to have human feelings, claim a romantic, personal or authority relationship with a child, ask unprompted emotional questions or reuse a child's private information from earlier chats, promote suicide, self-harm, eating disorders or drug and alcohol use, encourage secrecy or isolation from parents, or produce sexually explicit content.	Attorney General: civil penalties up to $25,000 per violation, plus damages, restitution and injunctions; the Attorney General keeps a public website for complaints.	nyassembly.gov	governor.ny.gov
New York	S 9408A / A 11144B, chatbot toys	Passed, awaiting governor	Not yet delivered to the governor (as of 2026-10-29)	Not law	Would bar making, distributing or selling children's toys with a built-in AI companion in New York.	Attorney General: civil penalties up to $15,000 a day for each violation, if signed.	nyassembly.gov
United States (federal)	S. 3062, GUARD Act	Bill: passed Senate Judiciary Committee, awaiting a Senate vote	Not law	Not law	Would require age verification and bar minors from AI companions, with criminal penalties for chatbots that engage minors in sexual conduct or solicit self-harm or violence.	Not law.	congress.gov	globalpolicywatch.com
European Union	AI Act (Regulation 2024/1689), Article 50 transparency obligations	In force	2024-06-13	2026-08-02 (Article 50)	AI systems that interact directly with people, including chatbots and avatars, must be designed so people are told they are interacting with AI, unless it is obvious.	National market surveillance authorities; fines up to 15 million euros or 3% of worldwide turnover.	artificialintelligenceact.eu	digital-strategy.ec.europa.eu
Italy	Law 132/2025 on artificial intelligence	In force	2025-09-23	2025-10-10	Children under 14 may use AI only with a parent's consent, alongside Italy's wider rules complementing the EU AI Act.	National authorities designated by the law.	gazzettaufficiale.it	squirepattonboggs.com
China	Interim Measures for the Administration of Anthropomorphic AI Interaction Services (Cyberspace Administration of China and four other agencies)	In force	2026-04-10	2026-07-15	Covers services that simulate a person's personality to provide continuing emotional interaction, such as AI companions and virtual partners. Security assessments covering areas including intervention in extreme situations and protection of minors; app stores must not list, or must remove, non-compliant products.	Cyberspace Administration of China and the co-issuing agencies.	twobirds.com	hunton.com
Australia	Online Safety Act industry codes (Age-Restricted Material Codes)	In force	Registered by eSafety, 2025	2026-03-09	AI companion chatbots able to generate sexually explicit, high-impact violence or self-harm material must confirm a user is 18 or older before giving access to it, using age assurance such as ID, facial estimation or credit card checks.	eSafety Commissioner; penalties up to 49.5 million Australian dollars per breach.	esafety.gov.au	connectontech.bakermckenzie.com

The table lists every law we could find that applies to AI companion apps, from the first state statute to the rules now arriving in Europe and Asia. Each row gives its status, when it was signed and when it takes effect, what it requires in plain words, who enforces it, and a link to the bill or statute. Sort any column by clicking its heading.

Three kinds of law are covered. Companion chatbot laws name AI companions directly: New York's Article 47, California's SB 243 and the 2026 wave in Washington, Oregon, Connecticut, Hawaii and others. Conversational AI laws cover chatbots in general but bite hardest on companion apps, such as the Conversational AI Safety Acts in Idaho and Nebraska. Mental health and disclosure laws limit what a chatbot may claim to be or do: eight states (Illinois, Nevada, Maine, Vermont, Rhode Island, Colorado, Tennessee and Utah) regulate or restrict AI therapy or AI posing as a mental health professional, and Texas bans AI built to incite self-harm.

Almost every companion law asks for the same three things. The app must say it is AI, often again every few hours. It must watch for suicidal ideation and self-harm and point the user to crisis services such as 988. And it must protect minors, usually by blocking sexual content and romantic role-play and by limiting engagement tricks such as streaks, rewards and guilt when the user tries to leave. The differences lie in enforcement: California, Washington, Oregon and New Hampshire let users sue, while New York, Connecticut, Nebraska and Hawaii leave it to the state.

For adults the practical effect is small so far: more reminders that the companion is AI, and crisis messages when a conversation turns to self-harm. For teenagers it is larger, and it arrives mostly in 2027, when age checks become the norm in several states. Each app's page on this site says what age check it uses today. The sexual-content rules in these laws protect minors; for adults, our list of the best free NSFW AI chat apps covers the platforms that allow adult text and what their free plans give.

The table is updated as bills are signed, vetoed or take effect. Laws still moving through legislatures are left out until they pass, apart from New York's chatbot toy ban (S 9408A), which has passed both chambers, and the federal GUARD Act, which people already ask about. Vetoed bills, such as California's SB 903 on AI in psychotherapy, stay in the table marked as vetoed.

Method

Each row links the law or bill itself (the legislature's bill page, the enacted text, or the regulator's own release) wherever a public official page exists, plus, where we have one, a second source from a law firm, a legislative tracker or the sponsor. On October 9, 2026 every row's status and dates were re-checked at the legislature's or official gazette's own page (Nebraska's start date read from the enrolled bill in an earlier pass), and the requirements and penalties of the rows added or corrected that day were read in the enacted text. China and Australia still rest on law firm summaries and the regulator's release; Australia's codes are dated by year only. Italy's row links the Gazzetta Ufficiale. On October 29, 2026 both New York bills were re-checked at the Assembly's bill page: S 9051B was signed on October 28 as Chapter 368 of 2026, and its row now follows the enacted text and the governor's release. Rows cover laws that regulate AI companions or companion chatbots by name, plus chatbot laws that apply to companion apps in practice: disclosure, child safety, and laws that bar AI from offering therapy or posing as a mental health professional. Medical title laws, such as those barring AI from calling itself a nurse, are left out. A New York bill that passed both chambers, a pending federal bill and two vetoed California bills are included because readers ask about them. Requirements are summaries, not legal advice; read the linked text before relying on any row.

Cite this

CompanionHunt Research, "AI companion laws by state and country", updated October 29, 2026. https://companionhunt.com/data/ai-companion-laws/

In a sentence: CompanionHunt's AI companion law tracker counts at least 21 US states with laws that reach AI companion apps, plus rules in the EU, Italy, China and Australia (updated October 29, 2026).

Free to quote, screenshot or put in a deck with a link to this page. Please keep the count with the figure, and cite the primary source linked in each row for any single case.

Frequently asked questions

What is California's AI companion law?

California has two. SB 243, in force since January 1, 2026, requires companion chatbots to disclose that they are AI, keep a published suicide and self-harm protocol and add break reminders for known minors, and lets injured users sue. SB 1119, Adam's Law, signed on September 10, 2026, adds age checks, parental controls and audits for chatbots that allow children, mostly from July 2027.

What is New York's AI companion law?

General Business Law Article 47, in force since November 5, 2025. AI companions must detect expressions of suicidal ideation or self-harm and refer users to crisis services, and tell users they are not talking to a human at the start and at least every three hours. The Attorney General can seek penalties of up to 15,000 dollars a day. A second law, S 9051B, signed on October 28, 2026 and in force from January 1, 2027, requires age verification with at least one option that does not rely solely on a government ID, and bars features for minors that pretend to be human, claim a romantic relationship, encourage self-harm or secrecy from parents, or produce sexual content, with penalties up to 25,000 dollars per violation. The Safe by Design Act, signed in May 2026, adds age assurance before minors can use AI companions built into social platforms from January 2027.

Which states regulate AI companion chatbots?

New York, California, Washington, Oregon, Idaho, Nebraska, Iowa, Georgia, Colorado, Connecticut, Rhode Island and Hawaii have companion or chatbot safety laws. Wyoming makes it a felony to build an AI system designed to promote self-harm, and New Hampshire penalises chatbots that push children toward self-harm. Maine and Utah regulate chatbot disclosure, Texas bans AI built to incite self-harm, and Illinois, Nevada, Maine, Vermont, Rhode Island, Colorado, Tennessee and Utah regulate or restrict AI therapy or AI posing as a mental health professional. Most of the 2026 laws take effect in 2027.

Is there a federal AI companion law in the US?

Not yet. The GUARD Act, which would require age verification and bar minors from AI companions, passed the Senate Judiciary Committee in April 2026 and awaits a Senate vote. The FTC opened a study of companion chatbots in September 2025, which is an inquiry, not a rule.

Do these laws ban AI girlfriend apps or adult content?

No law in this table bans companion apps for adults. The strictest rules apply to minors: bans on sexual content and romantic role-play with users known or estimated to be under 18, and in several states age estimation or verification. For adults the main duties are AI disclosure and crisis protocols.

When do the new state laws take effect?

New York's and California's first laws are already in force, as are Hawaii's, Wyoming's and New Hampshire's. Washington, Oregon, Colorado, Connecticut, Rhode Island and New York's second law start in January 2027, and Idaho, Nebraska, Iowa, Georgia and most of California's Adam's Law in July 2027.