Vermont Enacts H.816 Restricting AI Mental Health Therapy (2026)
Vermont Enacts H.816, Restricting AI From Providing Mental Health Therapy
Vermont Governor Phil Scott signed H.816 on June 17, 2026, barring companies from offering AI-delivered mental health therapy and joining a small group of states that began restricting AI therapists in 2025. The law lets clinicians use AI for administrative tasks but reserves diagnosis, treatment, and therapeutic decisions for a qualified mental health professional or an approved research study.
Information last verified on June 19, 2026. This is a developing story; we update it as the record changes.
Status: Signed into law by Governor Phil Scott on June 17, 2026, and effective on passage. Vermont follows Illinois, Nevada, and Utah, which each enacted restrictions on AI in mental health care during 2025.
Jurisdiction scope: This article addresses Vermont's H.816 and its limits on AI in the provision of mental health services in Vermont. It does not address other states' AI or telehealth rules, and it is not individualized advice. For the broader state framework, see Vermont AI laws and AI laws across the United States.
What Happened
On June 17, 2026, Governor Phil Scott signed H.816, An act relating to regulating the use of artificial intelligence in the provision of mental health services. The Governor's office listed the bill among those signed that day, alongside several unrelated measures, while vetoing two other bills. H.816 takes effect on passage, so its restrictions are operative now.
The act states its purpose plainly. According to the legislation, the goal is "to safeguard individuals seeking mental health services in Vermont from psychological harm, including death by suicide, by ensuring that these services are delivered by mental health professionals and not independently by artificial intelligence systems." That framing, suicide risk from chatbots acting as stand-in therapists, drove the bill through the House and Senate, which approved a conference version before final passage.
The core rule is a prohibition. A corporation or other entity may not offer or advertise mental health services, defined as counseling, therapy, or psychotherapy used to diagnose or treat a person's mental or behavioral health or to provide ongoing recovery support, when those services are delivered by an artificial intelligence system, unless the services are provided by a qualified mental health professional or are furnished as part of an approved research study. The definition reaches AI that makes therapeutic decisions, issues direct therapeutic communications, generates treatment plans or recommendations, or detects or interprets emotion or mental states. The law does not outlaw every use of AI in a clinical setting. It permits AI for administrative and supplementary functions, such as scheduling, billing, and transcription, so long as a qualified professional retains responsibility for the clinical work.
What the Law Actually Says
H.816 works through Vermont's professional-licensing and consumer-protection systems rather than through a new data-privacy framework. For a licensed mental health professional, using AI to make therapeutic decisions independently is treated as unprofessional conduct, the same category of violation that exposes a clinician to discipline by the relevant licensing board. For the companies that market AI tools to the public, the act reaches the act of offering or advertising AI-delivered mental health services in Vermont.
Enforcement runs through the Vermont Consumer Protection Act, 9 V.S.A. ch. 63, which prohibits unfair and deceptive acts in commerce and is enforced by the Vermont Attorney General. Tying the prohibition to that statute matters because the Consumer Protection Act already carries investigative and penalty tools, so the state did not have to build a new enforcement regime from scratch. The practical effect is that an AI product pitched to Vermonters as a therapist, rather than as a wellness or journaling aid, can be challenged as an unfair or deceptive practice.
The line the statute draws is between substituting for a clinician and supporting one. An AI scheduling assistant, a billing tool, or a transcription service that a licensed provider supervises sits on the permitted side. A chatbot that diagnoses a condition, recommends a treatment plan, or carries on therapeutic conversations on its own sits on the prohibited side. The distinction is also why industry and telehealth groups, including the Computer and Communications Industry Association, the Software and Information Industry Association, and the American Telemedicine Association, asked the Governor to veto the bill: they argued that the definitions of mental health services and therapeutic communication were broad enough to capture general wellness apps, peer-support features, and crisis tools that were never meant to act as licensed care. The Governor signed it without that change.
Analysis: Why This Matters
The following is analysis from the Recording Law Editorial Team.
H.816 is notable less for any single provision than for the regulatory route it chose. Rather than create a new AI-specific agency or a novel cause of action, Vermont folded AI therapy into two systems that already exist: professional licensing and consumer protection. That choice makes the law easier to enforce with current tools and harder to dismiss as symbolic. It also reflects a recurring pattern in state AI legislation, where lawmakers attach new conduct to familiar enforcement statutes instead of writing standalone penalty schemes.
The bill arrives during a national wave of attention to companion and therapy chatbots. Illinois, Nevada, and Utah each enacted restrictions on AI in mental health care during 2025, and similar measures have been introduced in numerous other states. Vermont's version is comparatively direct because it targets the provision of a regulated service, mental health care, instead of trying to regulate AI in the abstract. That framing gives it a clear hook, but it is also the source of the industry objection. The contested question, which the act's broad definitions leave open, is where a supportive wellness tool ends and an unlicensed therapist begins. How Vermont's Attorney General draws that line in any enforcement action will tell developers more about the law's reach than the statutory text does on its own.
We are not predicting how the law will be enforced or how any challenge to it would resolve. The narrower point is that Vermont has converted a widely shared worry, AI systems acting as unsupervised therapists, into an operative legal duty backed by an existing enforcement statute, and other states considering chatbot bills now have a concrete model to copy or to avoid.
How This Affects You
For Vermonters, the immediate effect is on what can be marketed to you as mental health care. A product offered as a therapist or as a substitute for clinical treatment now has to be backed by a qualified professional or an approved study. General wellness, meditation, and journaling apps that do not hold themselves out as providing mental health services are not the target of the prohibition, although the exact boundary is one the state will have to apply case by case.
For clinicians and the companies that sell tools into Vermont, the takeaway is about labeling and supervision. Using AI to handle scheduling, billing, or transcription under professional oversight is contemplated by the law; presenting AI as the entity that diagnoses or treats is not. If you operate in other states, do not assume Vermont's rule applies elsewhere, because most states have not enacted a comparable restriction and the bills under consideration differ in scope. None of this is legal advice about a specific product, practice, or situation, and the law in this area is moving quickly.
This is general legal information, not legal advice. It covers Vermont's H.816 and its restrictions on AI in the provision of mental health services, and it reflects sources verified on June 19, 2026. Laws change and this story is developing; consult a lawyer licensed in your jurisdiction about your specific situation.
Sources
- H.816, An act relating to regulating the use of artificial intelligence in the provision of mental health services (As Passed by Both House and Senate, Official): https://legislature.vermont.gov/Documents/2026/Docs/BILLS/H-0816/H-0816%20As%20Passed%20by%20Both%20House%20and%20Senate%20Official.pdf
- Bill Status, H.816 (2026 session), Vermont General Assembly: https://legislature.vermont.gov/bill/status/2026/H.816
- Office of Governor Phil Scott, "Action Taken by Governor Phil Scott on Legislation - June 17, 2026": https://governor.vermont.gov/press-release/action-taken-governor-phil-scott-legislation-june-17-2026
- Vermont Consumer Protection Act, 9 V.S.A. ch. 63: https://legislature.vermont.gov/statutes/chapter/09/063
Related articles
- Vermont AI laws
- AI laws across the United States
- Vermont's Data Privacy and Online Surveillance Act
- Louisiana's AI medical-recording disclosure law
Last updated: 2026-06-19. This is a developing story; details verified as of 2026-06-19.
Frequently Asked Questions
What did Vermont's H.816 do?
It bars corporations and entities from offering or advertising mental health services delivered by artificial intelligence unless the services come from a qualified mental health professional or an approved research study. Governor Phil Scott signed it on June 17, 2026, and it took effect on passage.
Does the law ban all AI in therapy offices?
No. H.816 allows AI for administrative and supplementary functions such as scheduling, billing, and transcription, as long as a qualified mental health professional retains clinical responsibility. It targets AI that independently diagnoses, treats, or conducts therapeutic communications.
How is the law enforced?
For a licensed professional, prohibited AI use is treated as unprofessional conduct subject to licensing-board discipline. More broadly, violations are enforceable under the Vermont Consumer Protection Act (9 V.S.A. ch. 63), which the Vermont Attorney General enforces against unfair and deceptive practices.
Is an AI wellness or meditation app illegal in Vermont now?
The prohibition targets services offered or advertised as mental health care, meaning counseling, therapy, or psychotherapy that diagnoses, treats, or supports recovery, including AI that makes therapeutic decisions, generates treatment plans, or interprets a user's emotions. General wellness tools that do not hold themselves out as providing mental health services are not the focus, although the act's broad definitions leave the exact boundary to be applied case by case.
Why did technology groups oppose H.816?
Groups including the Computer and Communications Industry Association and the Software and Information Industry Association urged a veto, arguing the definitions of mental health services and therapeutic communication were broad enough to capture general wellness and support tools. The Governor signed the bill without narrowing those definitions.
Does H.816 change my data-privacy rights?
Not directly. H.816 regulates who may provide AI-delivered mental health services, not consumer data rights. Vermont's separate data-privacy statutes address how personal data is collected and used, so the two operate alongside each other.
Sources and References
- H.816, An act relating to regulating the use of artificial intelligence in the provision of mental health services (As Passed by Both House and Senate, Official)(legislature.vermont.gov)
- Bill Status, H.816 (2026 session), Vermont General Assembly(legislature.vermont.gov)
- Office of Governor Phil Scott, Action Taken on Legislation - June 17, 2026(governor.vermont.gov)
- Vermont Consumer Protection Act, 9 V.S.A. ch. 63(legislature.vermont.gov)