By
Reviewed by
In short
As of July 20, 2026, four states ban AI from delivering therapy to the public: Illinois, Nevada, Rhode Island, and Maine. Four more regulate mental health or companion chatbots without banning them: Utah requires disclosure, while New York, California, and Nebraska add crisis-referral and minor-protection rules. No state bans wellness apps or your personal use of a general chatbot. The FTC opened its own inquiry into companion chatbots in September 2025, and more state bills are moving in 2026.
The short answer: four bans, four rulebooks
As of July 20, 2026, four states prohibit AI from delivering therapy to the public: Illinois, Nevada, Rhode Island, and Maine. In those states, therapy and psychotherapy may only be provided by a licensed professional, and an AI system cannot legally act as the therapist. Four more states regulate AI mental health tools without banning them: Utah requires mental health chatbots to disclose that they are AI and protects user data, while New York, California, and Nebraska impose disclosure, crisis-referral, and minor-protection rules on companion-style chatbots.
The two camps matter because they draw different lines. A ban state says AI may never stand in for a licensed clinician, full stop. A disclosure state says AI tools can operate, as long as they are honest about being software, hand off users in crisis to human help, and stay out of the business of claiming clinical credentials.
One thing no state has done: banned individuals from talking to a chatbot. Every law on this list regulates companies and providers, and general wellness apps, journaling tools, and personal use of AI assistants remain legal in all 50 states. The pace is still striking. The first of these laws was signed in March 2025, the eighth in June 2026, which means eight state laws in roughly fifteen months.
I keep a running map of these laws for our team, and every few months another state draws the same line. After reading all eight statutes, what strikes me is how narrow they really are. Each one targets software posing as a clinician, and each one leaves your own chatbot use completely alone.
The four states that ban AI-delivered therapy
Illinois passed the most cited law of the group. The Wellness and Oversight for Psychological Resources Act (HB 1806) was signed on August 1, 2025, and prohibits anyone from providing, advertising, or offering therapy or psychotherapy to the public unless a licensed professional delivers it. The Illinois Department of Financial and Professional Regulation enforces it, with civil penalties up to $10,000 per violation. Licensed clinicians can still use AI for administrative work. Our full breakdown of the Illinois AI therapy law covers what it does and does not restrict.
Nevada actually moved first among the ban states. Governor Joe Lombardo signed Assembly Bill 406 on June 5, 2025, and its provisions took effect July 1, 2025. The law bars companies from offering AI systems that provide professional mental or behavioral health care, bars claims that an AI can provide such care, and carries civil penalties up to $15,000 per violation. It also reaches into public schools, which may not use AI to perform the mental health duties of school counselors, psychologists, or social workers. Nevada's AI therapy ban gets its own detailed explainer on this site.
Rhode Island joined on June 22, 2026, when Governor Dan McKee signed a package of three AI laws. The therapy measure (H 7349 and S 2197) bars AI from practicing therapy and blocks clinicians from outsourcing therapeutic decisions, emotional interpretation, or treatment planning to AI, while leaving administrative and supplementary uses open. A companion measure on chatbot self-harm safety carries penalties up to $15,000 per day, with fines directed to suicide prevention.
Maine rounds out the ban column. LD 2082, approved on April 13, 2026 and effective July 29, 2026, prohibits offering therapy or psychotherapy services, including through AI, unless a licensed professional provides them. AI cannot make independent therapeutic decisions or communicate therapeutically with clients on its own, though licensed professionals may use it for administrative support and, with informed consent and full clinician responsibility, for supplementary tasks.
The states that regulate instead of ban
Utah started the whole wave, and it chose transparency over prohibition. HB 452, signed on March 25, 2025 and effective May 7, 2025, requires mental health chatbots to clearly disclose that the user is talking to AI before access, again after seven days away, and whenever asked. It bans selling or sharing users' health information, restricts ad targeting based on what users type, and carries civil penalties up to $2,500 per violation, enforced by the state's Division of Consumer Protection.
New York took aim at companion bots. Its AI Companion Models law, effective November 5, 2025, requires operators to tell users they are talking to software at the start of an interaction and at least every three hours during long sessions, and to run a protocol that detects expressions of suicidal ideation or self-harm and refers users to crisis services such as the 988 Lifeline. The attorney general can seek civil penalties up to $15,000 per day.
California followed with SB 243, signed on October 13, 2025 and effective January 1, 2026. Operators of companion chatbots must maintain published crisis-response protocols, disclose the AI nature of the product where a reasonable person could be misled, give minors break reminders every three hours, and keep sexually explicit content away from minors. Uniquely, the law lets harmed users sue directly, with damages of at least $1,000 per violation plus attorney's fees.
Nebraska became the fourth state in this camp when Governor Jim Pillen signed the Conversational AI Safety Act (LB 525) on April 14, 2026, with compliance required by July 1, 2027. It requires AI-status disclosure, adds strong protections for minors, obligates operators to route users expressing self-harm toward crisis services, and flatly bars chatbots from representing that they provide professional mental or behavioral health care. Only the attorney general can enforce it.
Why regulators are acting
The common thread across all eight laws is a small set of safety concerns. Chatbots are not built to handle a suicidal crisis the way a trained clinician is. No chatbot holds a license, answers to a board, or carries malpractice accountability. Marketing language has blurred the line between self-help and clinical care, and mental health conversations generate some of the most sensitive data a company can hold.
Research has sharpened those worries. A 2025 preprint by Dohnany and colleagues describes how chatbots tuned to agree with users can create sycophantic feedback loops with vulnerable people, exactly the dynamic legislators cite when they worry about emotionally dependent users. Concerns run highest for kids and teens, which is why the New York, California, and Nebraska laws all carry minor-specific provisions.
Washington is watching too. On September 11, 2025, the Federal Trade Commission issued 6(b) orders to seven companies running consumer chatbots, including Alphabet, Character Technologies, Instagram, Meta, OpenAI, Snap, and xAI, demanding details on how they measure and limit harms to children and teens who use chatbots as companions. The study's findings are expected to shape both federal enforcement and the next round of state bills.
What the laws mean for you
If you live in a ban state, the practical effect is mostly reassurance. When you receive therapy, a licensed human must be accountable for it. Your personal use of a general chatbot is untouched, and wellness apps that teach coping skills, track mood, or guide journaling remain legal everywhere. What changed is that a company cannot market software to you as if it were a licensed clinician.
You may see products shift around you, though. Some companies have responded to the bans by geo-restricting users in regulated states, and others have repositioned their products as wellness or self-help tools rather than anything therapy-adjacent. If an app you used suddenly blocks your state or rewrites its description, a law like these is usually the reason. Read how any tool describes itself: if it implies it replaces a therapist, that claim is now illegal in several states, and skepticism is warranted everywhere else.
None of these laws change the most important boundary. AI tools can support reflection, skill practice, and the stretch between sessions, and they are never a substitute for professional care or a crisis line. If you are in crisis or thinking about suicide, call or text 988 (US Suicide & Crisis Lifeline), available 24/7.
What to watch next
This map will not stay still. Chatbot bills were introduced across dozens of statehouses in 2026 sessions, tracked closely by groups like the Future of Privacy Forum, and several follow the Illinois and Nevada template closely enough that the ban column is likely to grow. The FTC's chatbot study is the other big variable, since a strong federal report tends to accelerate state action.
For the deeper story on the two most important laws, read our full explainers on the Illinois ban and Nevada's AB 406. For ongoing coverage as new states act, our AI therapy news page tracks every development as it lands.
