Regulon.AI — VA-HB-668 · VA HB 668 (Mental Health AI Restrictions)

2026-01-13 introduced Prefiled and ordered printed; Offered 01-14-2026 26104644D

2026-01-13 committee referral Referred to Committee on Communications, Technology and Innovation

2026-01-20 other Fiscal Impact Statement from Department of Planning and Budget (HB668)

2026-01-29 other Assigned HST sub: Technology and Innovation

2026-02-04 other House subcommittee offered

2026-02-04 other Subcommittee recommends laying on the table (6-Y 4-N)

2026-02-18 other Left in Committee Communications, Technology and Innovation

Summary

Regulates the use of artificial intelligence systems by licensed mental health service providers in Virginia. Mental health service providers must maintain full responsibility for all AI interactions, outputs, and data use. AI systems may not independently make therapeutic decisions, directly interact with clients in therapeutic communication without provider oversight, or generate treatment plans without licensed professional review and approval. Recording or transcribing sessions with AI requires written disclosure and patient consent. No person or entity may offer therapy or counseling services through AI unless conducted by a licensed mental health service provider. Violations are subject to civil penalties up to $10,000 collected by the Department of Health Professions.

Enforcement & Penalties

Enforcement Authority

Civil penalty collected by the Department of Health Professions. No private right of action is created. Enforcement is agency-initiated through the Department of Health Professions.

Penalties

Civil penalty not to exceed $10,000 per violation. Penalty is deposited into the general fund. No private damages, injunctive relief, or attorney fee provisions.

Who Is Covered

"Mental health service provider" has the same meaning as provided in § 54.1-2400.1.

What Is Covered

"Artificial intelligence system" means any machine learning-based system that, for any explicit or implicit objective, infers from the inputs such system receives how to generate outputs, including content, decisions, predictions, and recommendations, that can influence physical or virtual environments. "Artificial intelligence system" does not include any artificial intelligence system or general purpose artificial intelligence model that is used for development, prototyping, and research activities before such artificial intelligence system or general purpose artificial intelligence model is made available to deployers or consumers.

Compliance Obligations 5 obligations · click obligation ID to open requirement page

HC-02 AI in Licensed Professional Practice Restrictions · HC-02.1 · Professional · Healthcare

§ 54.1-2400.1:1(B)

Plain Language

Mental health service providers may use AI systems to assist in therapy or counseling, but only if the provider maintains full responsibility for all interactions, outputs, and data use associated with the system. This is a gating condition — AI use in therapy is permissible only under this supervisory framework. The provider cannot delegate professional responsibility to the AI system.

Statutory Text

B. A mental health service provider may use an artificial intelligence system to assist in providing therapy or counseling services if such mental health service provider maintains full responsibility for all interactions, outputs, and data use associated with the system.

HC-02 AI in Licensed Professional Practice Restrictions · HC-02.4 · Professional · Healthcare

§ 54.1-2400.1:1(B)(1)-(2)

Plain Language

When a mental health service provider uses AI to record or transcribe a therapy or counseling session, two requirements must be met: (1) the provider must disclose to the patient (or their legally authorized representative) that AI will be used and explain the specific purpose; and (2) at the initial appointment, the provider must disclose its AI use policies and obtain written or digital consent. If AI policies change after consent is obtained, the provider must notify all patients who previously consented. Unlike CA SB 243's 24-hour advance consent requirement, this bill requires disclosure and consent at the initial appointment with no minimum advance notice period. Consent is revocable by implication of the consent framework but no explicit revocability language appears.

Statutory Text

No licensed professional shall be permitted to use an artificial intelligence system in providing therapy or counseling services pursuant to this subsection when the session is recorded or transcribed unless: 1. The mental health service provider discloses to the patient or the patient's legally authorized representative (i) that an artificial intelligence system will be used and (ii) the specific purpose of the artificial intelligence system that will be used; and 2. At the initial appointment, the mental health service provider discloses its artificial intelligence system use and policies related to such use and the patient or the patient's legally authorized representative provides written or digital consent to the use of an artificial intelligence system as permitted by this section. The mental health service provider shall give notice of any change in its policies related to the use of an artificial intelligence system to any patient or such patient's legally authorized representative who has consented to the use of an artificial intelligence system pursuant to this subdivision.

HC-02 AI in Licensed Professional Practice Restrictions · HC-02.3 · ProfessionalDeployer · Healthcare

§ 54.1-2400.1:1(C)

Plain Language

No person or business entity may offer, advertise, or provide therapy or counseling services — including through an AI system — to the public in Virginia unless those services are conducted by a licensed mental health service provider. This effectively prohibits standalone AI therapy products that operate without a licensed provider. It also prohibits advertising AI-only therapy or counseling services. The prohibition applies to any person or business entity, not just licensed providers — it catches AI companies and platforms that might attempt to offer unlicensed AI therapy directly to consumers.

Statutory Text

C. A person or business entity may not provide, advertise, or otherwise offer therapy or counseling services, including through the use of an artificial intelligence system, to the public in the Commonwealth unless the therapy or counseling services are conducted by a mental health service provider.

HC-02 AI in Licensed Professional Practice Restrictions · HC-02.2 · Professional · Healthcare

§ 54.1-2400.1:1(D)(1)-(3)

Plain Language

Mental health service providers face three categorical prohibitions on AI use: (1) AI may not make independent therapeutic decisions — every therapeutic decision requires professional judgment; (2) AI may not directly interact with clients in therapeutic communication without provider oversight — the definition of therapeutic communication is broad, covering clinical interactions, emotional support, guidance, treatment plan collaboration, and behavioral feedback; and (3) AI may not generate therapeutic recommendations, diagnose, or implement treatment plans without the licensed professional's review, oversight, and approval. AI may still be used for administrative support tasks that do not involve therapeutic communication (e.g., scheduling, billing, logistics communications).

Statutory Text

D. A mental health service provider may use an artificial intelligence system only to the extent that such use meets the requirements of subsection B. No mental health service provider shall allow an artificial intelligence system to: 1. Make independent therapeutic decisions; 2. Directly interact with clients in any form of therapeutic communication without provider oversight; or 3. Generate therapeutic recommendations, diagnose, or implement treatment plans without review, oversight, and approval by the licensed professional.

HC-02 AI in Licensed Professional Practice Restrictions · Professional · Healthcare

§ 54.1-2400.1:1(E)

Plain Language

All records maintained by mental health service providers — including records generated through or involving AI systems — and all communications between patients and providers must remain confidential under Virginia's existing health records privacy law (§ 32.1-127.1:03). Disclosure is permitted only as allowed under that statute. This extends existing health privacy protections to AI-involved records and communications, ensuring that AI-generated transcripts, session recordings, and related data receive the same confidentiality protections as traditional clinical records.

Statutory Text

E. All records kept by a mental health service provider and all communications between an individual seeking therapy or counseling services and a mental health service provider shall be confidential pursuant to the requirements of § 32.1-127.1:03 and shall not be disclosed unless such disclosure complies with the requirements of § 32.1-127.1:03.