FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Counsellors and psychotherapists
An AI risk score can prompt a therapist to look again, but it cannot replace their own suicide risk assessment, and BACP members must not defer that decision to AI. A missed risk is among the most serious claims a therapist can face, and an automated score that was relied on without question will be hard to defend.
Part of: AI in counselling and psychotherapy
In short
Some online platforms and services use AI to flag clients who may be at risk, for example from language in messages, questionnaire scores or session patterns. BACP 2.1(e)(v) says members must not defer or outsource decision-making to AI and remain responsible for critically evaluating its output. Data protection law adds limits: DUAA 2025 s80 restricts solely automated decisions with significant effects based on special category data such as health. A low score is not reassurance and a high score is not a diagnosis. PI usually responds to claims about missed risk, subject to the terms.
Therapists meet AI risk tools in several settings. Online therapy platforms may scan messages for words linked to self-harm and alert the practitioner. Services may score intake questionnaires automatically and route higher scores to faster appointments. Some practice software highlights changes in outcome measures between sessions. In each case the tool produces a flag or a score; what happens next depends on the people involved.
The main duties that apply are:
BACP's Ethics essentials session describes the practitioner as the "ethical gate". With risk, that gate is your clinical judgement.
Section 80 of the Data (Use and Access) Act 2025 inserted Articles 22A to 22D into UK GDPR from 5 February 2026. They restrict decisions based solely on automated processing that have significant effects, with tighter limits where special category data such as health information is involved. A service that refuses, delays or prioritises therapy purely on an AI risk score, with no meaningful human review, is the kind of process these rules target. Individual practitioners rarely design these systems, but you should know whether a human is genuinely reviewing decisions on your clients, and whether that human is you.
Illustrative scenario: a therapist works through a platform that shows a green, amber or red risk indicator next to each client. A client who has been green for weeks mentions, briefly and calmly, that they have been giving away belongings. The indicator stays green, and the therapist, reassured, does not explore it. The client later attempts suicide. In the complaint and claim that follow, the question is not what the indicator said, but whether a competent therapist would have assessed the risk, and whether the therapist let the tool make that judgement.
The reverse also causes harm: a false red flag that triggers an unnecessary safeguarding referral can damage trust and lead to a complaint.
Claims alleging failure to assess or act on suicide risk are serious and are what therapist PI and liability cover is usually bought for, subject to the policy terms. Check how your policy responds to injury or death claims (some PI wordings carry a bodily injury exclusion, while therapist malpractice policies may be built to include it), whether platform work is within your declared activities, and whether the platform's own insurance is relevant. Notify incidents promptly; see circumstance notification.
Useful controls include:
Apex arranges PI for counsellors and psychotherapists, including those working through platforms and services. Related reading: AI on online therapy platforms, AI triage in EAP work and the counselling AI hub.
No. Use it as a prompt, not a conclusion. BACP 2.1(e)(v) says members must not defer decisions to AI and remain responsible for evaluating its output. Carry out and record your own assessment, and note any disagreement with the tool.
Solely automated decisions with significant effects that use health data are restricted under UK GDPR as amended by the Data (Use and Access) Act 2025, from 5 February 2026. Meaningful human review is the key safeguard. Services, not individual therapists, usually design these processes, but you should know how they work.
Yes. BACP 3.1(d) requires clients to be told about platforms or tools that may monitor or collect data, and 2.1(e)(iii) requires honesty about AI use. Check the platform explains risk scanning, and add your own explanation at contracting if it does not.
Therapist PI and liability cover is usually intended to respond to claims alleging failure to assess or manage risk, subject to the policy terms. Check any bodily injury wording, your declared activities, and notification conditions. A broker can confirm how your policy is set up.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
