Aipedia: Counsellors and psychotherapists

Using AI to prepare for clinical supervision

Aipedia, by the Apex broking team · Last reviewed 2026-10-30

You can use AI to organise your thoughts before supervision, but not to put identifiable client material into an untested tool or to do the reflecting for you. BACP treats supervision as the place for in-depth reflection, and an AI summary that leaks or distorts a client's story undermines both.

Part of: AI in counselling and psychotherapy

In short

Some therapists paste session notes into a generative AI chatbot to produce a case summary or list of themes for supervision. The BACP Ethical Framework 2026 requires regular, independent supervision (4.3) and collaboration with supervisors on ethical dilemmas (4.3(d)). It also requires a risk assessment and informed consent before client data goes into AI (2.1(e) and (iv)). BACP's guidance describes the practitioner as an "ethical gate" and says AI cannot do reflective practice for you. Removing names is often not enough to anonymise therapy material. A breach of confidence from this use is likely to be treated as a professional failing.

How therapists use AI before supervision

Supervision time is limited, and caseloads can be heavy. Some practitioners use an AI assistant to condense several weeks of notes into a one-page summary, to suggest themes or hypotheses, or to draft questions to take to their supervisor. Others use it to tidy reflective writing for accreditation portfolios.

Used on your own reflections, with no client information, this is similar to any writing aid. The difficulty starts when client material goes in, and when the AI's framing of the case starts to replace your own.

The duties that apply

These duties apply whether or not AI is involved:

4.3(a) to (d) of the BACP Ethical Framework 2026: supervision must be regular, independent of line management, with a competent supervisor, and you must collaborate with supervisors on ethical dilemmas and challenges. Whether and how you use AI is itself such a dilemma.
2.1(e): assess any AI tool before use and understand how it handles and stores data.
2.1(e)(iv): informed consent before entering a client's personal data. Supervision is a known limit of confidentiality; an AI tool is not, unless you have told the client.
2.1(e)(v): do not defer decisions to AI, including formulation and risk judgements.
3.1(a): the duty to keep information shared confidential.

BACP's Ethics essentials session on AI calls the practitioner the "adult in the room" and an "ethical gate", and says AI can help with admin but not with ethical reasoning or reflective practice.

The limits of anonymisation

Removing a client's name rarely makes therapy material anonymous. Age, occupation, family structure, a distinctive event or a small town can identify someone, especially in a close community. If the material can identify a person, it is still personal data, and usually special category health data. The safer approach is to write your own brief, high-level prompt without client detail, or to use AI only on your own reflective notes about your process.

What can go wrong

Illustrative scenario: a psychotherapist pastes three months of notes, with names removed, into a free chatbot to produce a supervision summary. The notes mention the client's employer, a recent court case and their child's school. The chatbot's terms allow inputs to be used for training. Later, the client learns of this through a data access request that lists the tools used, and complains that their story was shared with a technology company without consent. Separately, the supervisor notices the summary has framed the case around a theme the therapist had not considered, and that framing has quietly shaped the work.

How PI usually responds

A breach of confidence claim from a client is the kind of allegation therapist PI commonly covers, subject to the policy terms and any limits on data claims. Professional conduct complaints are a separate process; some policies include legal costs for defending them, some do not. Check whether your policy treats unauthorised disclosure through a third-party tool differently from other breaches, and read how PI treats GDPR and data breach claims. If a client raises a concern, consider whether it is a circumstance to notify.

Practical controls

You can reduce the risk with these steps:

Do not put identifiable client material into a general-purpose AI tool.
If you use AI at all, use it on your own process notes or a deliberately generic prompt.
Agree with your supervisor how AI may be used in preparing for sessions.
Bring your own formulation first, then test the AI's suggestions against it, not the other way round.
Record your risk assessment of the tool, and disclose its use in your privacy notice if any client data is involved.

How Apex helps

Apex arranges PI for counsellors and psychotherapists, and a broker can confirm how your policy handles confidentiality claims and complaint defence costs. Related guides: AI notetakers, AI use by students on placement and the counselling AI hub.

Sources

Frequently asked

Can I use AI to write my supervision notes?

You can use AI on your own reflections, but avoid entering identifiable client material. BACP requires a risk assessment of the tool and informed consent before any personal data goes in. AI should support your thinking, not replace the in-depth reflection supervision is for.

Is removing the client's name enough to anonymise notes?

Usually not. Details such as age, job, family, location or a distinctive event can identify a person. If the material could identify someone, it remains personal data, and therapy content is special category health data under UK GDPR.

Can my supervisor use AI on what I share?

The same duties apply to supervisors who are BACP members: assess the tool, protect confidentiality and get consent before entering personal data. Agree at the start of the supervisory relationship whether either of you uses AI, and how material shared in supervision is stored.

Should I tell my supervisor I use AI?

Yes. BACP 4.3(d) requires you to collaborate with supervisors on ethical dilemmas and challenges, and AI use is one. Discussing it also gives you a record that you thought the issue through, which helps if a complaint is raised later.

Check your PI covers how you use AI

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 callback

Apex 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.