HB 2599
In CommitteeHouse
Therapy and psychotherapy
Protecting the integrity of therapy and psychotherapy services.
This status may be delayed. See Action History below for the latest updates.
How does a bill become law?
- Introduced: The bill is filed and assigned a number.
- Committee: A subject-matter committee holds hearings, takes public testimony, and decides whether to advance the bill.
- Floor Vote: The full chamber (House or Senate) debates and votes on the bill.
- Opposite Chamber: The bill repeats the committee and floor vote process in the other chamber.
- Governor: The Governor reviews the bill and decides whether to sign or veto it.
- Signed: The bill has been signed into law.
AI Analysis
This bill bans AI systems from independently providing mental health therapy or psychotherapy in Washington, requiring that all such services be delivered or directly supervised by licensed professionals. It also mandates clear consent from clients before AI tools are used in therapy sessions and defines strict limits on how AI may assist—without replacing—human clinicians.
- Prohibits AI systems from making independent therapeutic decisions, directly interacting with clients, generating treatment plans without licensed professional review, or detecting client emotions or mental states.
- Requires licensed professionals to get written, informed consent from clients (or their legal representatives) before using AI tools that involve recording or transcribing therapy sessions.
- Defines and limits AI use to 'administrative support' (e.g., scheduling, billing) or 'supplementary support' (e.g., organizing resources, anonymized data analysis), always under the licensed professional’s full responsibility.
- Bars any unlicensed person or entity—including AI companies—from offering or advertising therapy or psychotherapy services in Washington, even online.
- Makes violations of these rules subject to discipline under the Uniform Disciplinary Act and treats them as unfair or deceptive practices under the Consumer Protection Act.
Who is affected
- Licensed mental health professionals (e.g., psychologists, counselors, social workers, psychiatrists) — Must ensure AI tools used in therapy are only for administrative or supplementary support, not for direct client interaction or independent clinical decisions; must obtain written consent before using AI in sessions involving recording or transcription.
- Clients receiving therapy or psychotherapy services — Must be informed in writing and give explicit, voluntary consent before AI tools are used in their therapy sessions; have the right to refuse such use at any time.
- AI companies and technology providers — Cannot offer or advertise therapy services using AI without a licensed professional directly overseeing and being responsible for all clinical interactions.
- Unlicensed entities or individuals offering mental health services — May face disciplinary action or legal penalties if they allow AI to make independent therapeutic decisions or interact directly with clients without proper oversight.
Pro/Con Analysis
Stronger case for benefits
Potential Benefits (5)
Prevents AI systems from making independent therapeutic decisions or directly interacting with clients, protecting vulnerable individuals—including minors and those in crisis—from potentially harmful, unvalidated, or algorithmically biased clinical guidance.
HealthcarePeopleRef: Sec. 2(1)(a)-(d)Requires explicit, written, revocable consent before using AI in session recording or transcription, reinforcing client autonomy and privacy rights in an era of pervasive digital surveillance and data misuse.
Rights & LibertiesPeopleRef: Sec. 2(2)(a)-(b)Bars unlicensed entities—including AI companies—from offering or advertising therapy services, reducing consumer confusion and preventing deceptive marketing that could lead individuals to forgo evidence-based care.
Public SafetyPeopleRef: Sec. 4Holds licensed professionals fully responsible for AI use in therapy and treats violations as unfair/deceptive practices under the Consumer Protection Act, strengthening accountability and consumer recourse.
Public SafetyPeopleRef: Sec. 2(1)(e), Sec. 5Prohibiting AI from detecting client emotions or mental states prevents misdiagnosis or overreach by tools that lack clinical validation—especially important for minors and those with complex or stigmatized conditions.
Public SafetyPeopleRef: Sec. 2(1)(d)
Potential Concerns (5)
Prohibiting AI from detecting client emotions or mental states may prevent access to emerging tools that could improve diagnostic accuracy or early intervention for conditions like depression or PTSD—particularly in underserved areas with therapist shortages.
HealthcareRef: Sec. 2(1)(d)Mandating written informed consent for AI-assisted session recording or transcription adds administrative burden and may delay or discourage use of helpful tools (e.g., transcription for note-taking), especially for clinicians in time-constrained or under-resourced settings.
HealthcareRef: Sec. 2(2)(a)(ii)Barring unlicensed entities—including AI companies—from offering or advertising therapy services may reduce innovation and limit consumer access to low-cost digital mental health tools that, while not clinical therapy, serve as important triage or self-help resources.
Business & EmploymentRef: Sec. 4Requiring licensed professionals to personally review and approve all AI-generated treatment plan components increases clinician workload and may slow care delivery, especially in high-volume or public mental health settings.
Business & EmploymentRef: Sec. 2(1)(c)Treating violations as consumer protection act violations exposes AI companies and startups to broad liability, potentially chilling investment in mental health tech—even for non-clinical tools that could expand access.
Business & EmploymentRef: Sec. 5
Who Is Most Affected
Clients benefit significantly from stronger protections against unvetted AI therapy, especially vulnerable populations like minors and those in crisis. Consent requirements enhance autonomy and trust in care.
Licensed professionals retain clinical authority and avoid liability for AI errors, but face added administrative duties (e.g., consent documentation) and may lose access to time-saving AI tools.
AI companies and tech developers face significant restrictions on product design and marketing; those offering standalone mental health tools (even non-clinical) may be barred from Washington’s market.
Unlicensed providers (e.g., wellness coaches, peer-run platforms) can no longer use AI to offer services that resemble therapy, reducing low-cost alternatives—even if they fall in a regulatory gray area.
State licensing boards gain clearer authority to discipline violations, but must invest in enforcement capacity and legal interpretation of AI-related misconduct.