Generative-AI therapy chatbots (Therabot; experimental)

Domain IX · Biomedical, neuro & digital · 2019 · Dartmouth

What it is

Generative-AI therapy chatbots are conversational programs built on large language models. They write new, open-ended replies rather than choosing from pre-written scripts. Rule-based bots such as Woebot and Wysa are covered in 096. General-purpose assistants (ChatGPT-type tools) used for emotional…

Format

• Who: adults recruited nationally through online advertising and screened into three groups: major depressive disorder, generalised anxiety disorder, or clinically high risk for feeding and eating disorders. The trial had 210 participants: 106 used Therabot and 104 were on a waitlist. About 27% of…
• Format: smartphone text conversation. There were 4 weeks of daily prompts to use it, then 4 weeks of continued access without prompts, so 8 weeks in total. Use averaged about 260 messages over about 24 days, and about 6 hours, in the first 4 weeks.

Who practises it, and how they are trained or regulated

• FDA (US), 2026: a revised General Wellness guidance (6 January 2026) describes enforcement discretion for low-risk wellness products, such as stress-management or relaxation claims. Disease-related claims can bring a product into device review. A revised Clinical Decision Support guidance (29 January 2026) says functions that give recommendations directly to patients or caregivers do not qualify for the CDS…
• US states, 2025 laws: Illinois, HB 1806 (Wellness and Oversight for Psychological Resources Act, Public Act 104-0054, signed August 2025; sources give 1 or 4 August). It bars offering AI-delivered therapy. In therapy, AI may not make independent therapeutic decisions, communicate therapeutically with clients directly, produce treatment plans without licensed review, or detect emotions or mental states.…
• California requires disclosure when generative AI writes clinical patient communications. Its Attorney General advised in January 2025 that AI cannot practise licensed professions.

Evidence status, as the guide reports it

• Age limits: the trial enrolled adults only. In October 2025, Character.AI stopped users under 18 from open-ended chats with its personas and introduced age verification that places users in age brackets. APA asks for child-specific safeguards and age-appropriate pre-deployment testing, and the FTC…

Questions to explore with your client

• What drew you to it, and what are you hoping it will change?
• Does it involve any substance, plant preparation, remedy or supplement? Has your doctor or prescriber been told, especially about any medication you take?
• Does it involve fasting, a special diet or 'detox'? Have you discussed this with your doctor?
• Is it a group, retreat or community? How do they respond if someone wants to leave or raises a concern?
• Does it use a device, heat, cold, water or a physical procedure? Has a medical professional checked that it is safe for you?
• Who runs it, and what training or registration do they hold?
• What actually happens in a session or meeting, and how often?
• What does it cost, and is there pressure to commit to more?
• How do you feel afterwards, in the hours and days that follow?
• Does it fit with, or clash with, the work we do here?

What practitioners of it watch for

• vendors marketing "AI therapist" add-ons to practices.
• the client feeling criticised for using it. Frame this as information, not a verdict.
• self-blame and re-exposure. Do not require the client to re-read harmful transcripts.
• rebound use and low mood. It is working if the client has more human contact and the crisis plan is in use.
• the client stopping human care to join. It is working if the client can explain the trial's safety arrangements.

When to involve someone else

• New or worsening physical symptoms, or advice to stop or change prescribed medication: involve the client's doctor or prescriber.
• Signs of coercion, isolation from family and friends, escalating costs or secrecy: discuss openly and consider safeguarding.
• A crisis or thoughts of suicide: follow your crisis protocol. In the US, call or text 988; in the UK and Ireland, Samaritans on 116 123.

Drawn automatically from the guide's own sections. Stay curious: neither endorse nor dismiss what the client values, and check anything that matters against the full guide and its sources. Not medical advice.