HB 668 Virginia House of Delegates · 2026 Regular Session

Mental health service providers; use of artificial intelligence system, civil penalty.

Summary

Use of artificial intelligence system by mental health service providers; civil penalty. Permits the use of an artificial intelligence system by mental health service providers for administrative support and supplementary support, as those terms are defined in the bill, and prohibits the use of an artificial intelligence system to provide therapy or counseling services without a mental health service provider. The bill specifies that records kept by mental health service providers must comply with health records privacy requirements; creates an exception for religious counseling, peer support, or self-help materials and educational resources; and establishes a civil penalty not to exceed $10,000 for violations of the statute.

Issues: ✓ Healthcare

Bill status in committee 1 of 4 stages cleared

Introduction

Jan 2026

Committee Review

Floor Vote

Governor

Introduced Jan 13, 2026 Last action Feb 18, 2026

Maddy AI version diff · 1 comparison

What changed between versions

Introduced → Communications, Technology and Innovation Substitute · 4 edits

MODERATE

The bill was amended by the House Committee to remove specific definitions for 'administrative support' and 'supplementary support,' broadening the scope to allow AI use in any non-therapeutic task. The committee also replaced the strict prohibition on AI detecting emotions with a general ban on independent therapeutic decisions and added a requirement for provider oversight when AI interacts with clients. Additionally, the required patient consent method was expanded to include digital consent options.

Scope change

The bill's scope was broadened by removing the specific list of allowed administrative tasks, meaning AI can now be used for any task that does not involve therapeutic communication, rather than just the three previously listed examples.

DEFINITION

Removed the specific list of examples for 'administrative support' and 'supplementary support,' allowing AI use in any non-therapeutic task.

REQUIREMENT

Expanded consent requirements to explicitly allow for digital consent in addition to written consent.

Replaced the specific ban on AI detecting emotions with a broader prohibition on AI making independent therapeutic decisions or interacting with clients without provider oversight.

Added a requirement that AI must not generate recommendations or treatment plans without review, oversight, and approval by a licensed professional.

Floor votes

How they voted

No floor votes recorded yet.

Full legislative history

Actions timeline

Total actions

Key actions

Committee

Feb 18, 2026

Lower · Passed

Left in Committee Communications, Technology and Innovation

lower

Feb 4, 2026

Lower · Passed

Subcommittee recommends laying on the table (6-Y 4-N)

lower

Feb 4, 2026

Lower · Passed

House subcommittee offered

lower

Jan 29, 2026

Committee

Assigned HST sub: Technology and Innovation

lower

Jan 13, 2026

Committee

Referred to Committee on Communications, Technology and Innovation

lower

1 primary · 7 co-sponsors