0 of 2 psychologist insurers ask about AI

Ask your professional indemnity insurer whether they need to know you use an AI note-taking tool, and the honest answer, checked directly this week against the two main routes to psychologist indemnity cover in the UK, is that neither policy's application process asks the question. That silence is not the same as permission. UK commercial insurance runs on a duty to disclose material facts whether or not the form asks for them, and the wider professional indemnity market has spent 2026 treating AI use as exactly that kind of material fact, for solicitors, for accountants, for technology businesses. Psychology hasn't caught up yet, and that gap is now something you have to manage yourself, not something your insurer will flag for you.

AI is already an underwriting question, just not yours yet

Professional Indemnity Insurance Brokers' July 2026 market analysis, published 9 July 2026 and updated 13 July 2026, is blunt about where the market has landed: "Several UK insurers and law firms published guidance in June warning that this is now a genuine Professional Indemnity exposure. Underwriters are beginning to ask insureds how AI is being used within their businesses and what controls exist." The trigger was a wave of lawyers relying on generative AI that fabricated case law, but the report frames the underwriting shift as cross-sector, listing the questions now circulating across renewals: does the policyholder use AI, who reviews AI output, are humans signing off the advice. It calls this "one of the biggest emerging PI underwriting themes of 2026."

Miller Insurance's 20 July 2026 guidance on demonstrating responsible AI use to insurers goes further into what a well-prepared answer looks like, even though it's written for law firm clients. It lists five things underwriters want addressed: the scope and purpose of AI in the practice, governance and oversight including a human reviewing important outputs before they go out, data privacy and where the data is processed, staff training on the tool's risks, and the limits of liability in any contract with the AI vendor. None of that is specific to law. A psychology practice using an AI scribe has the same five answers to give, it just hasn't been asked for them yet.

So what for you: the questions an underwriter would ask if your renewal form caught up are already public. You can answer them now, in writing, before anyone asks, rather than scrambling to reconstruct the answer after a claim.

The duty that doesn't wait for the form to ask

An independent psychology practice buying professional indemnity cover for its business, whether as a sole trader, a partnership or a limited company, is buying non-consumer insurance. That distinction matters because non-consumer contracts sit under Part 2 of the Insurance Act 2015, not the separate consumer disclosure rules that apply to personal policies like home or travel insurance. The Act's explanatory notes, and the legal summaries written around it, describe the duty it creates as a duty of "fair presentation of the risk": disclosure of every material circumstance the insured knows or ought to know, or enough information to put a prudent insurer on notice and prompt further questions. A presentation is only fair if it's reasonably clear and accessible, and the facts in it are substantially correct.

The consequence of getting this wrong isn't automatically the end of your cover. The Act introduced proportionate remedies for most breaches, meaning an insurer may reduce a payout rather than refuse it outright, reserving full avoidance of the policy for breaches that were deliberate or reckless. But "proportionate" still means a live claim can come back smaller than expected, or contested, precisely at the point you need it most, and that risk exists whether or not your insurer's proposal form happened to ask about AI.

So what for you: the fact that your renewal form has no AI question is not legal cover for staying quiet about AI use. If a tool you use is a material circumstance a prudent insurer would want to know about, and an AI scribe touching every clinical record plausibly is, the duty to volunteer it sits with you, not with whoever wrote the form.

What the two main policies actually say about AI, checked this week

Most independent UK psychologists reach professional indemnity cover through one of two routes carrying a British Psychological Society or Health and Care Professions Council discount: Marsh Commercial's Oxygen scheme, underwritten by Hiscox, or Everywhen (formerly Towergate). Both were checked directly on 5 September 2026. Oxygen's psychologist page lists its standard cover, professional indemnity and malpractice, complaints and disciplinary hearings, confidential notes and records, crisis containment, and its full FAQ section, covering everything from run-off cover to what counts as a material change mid-policy, contains no mention of AI anywhere. Everywhen's psychologist page is the same shape: a detailed key-exclusions list naming fines and penalties, employee injury claims, and liabilities better suited to other policy types, but nothing about AI, either as an exclusion or as an acknowledged part of modern practice.

The gap is more pointed than it first looks, because Hiscox, the underwriter sitting behind the Oxygen scheme, has already solved this problem elsewhere in its own book. In May 2025, Hiscox rewrote its Technology Professional Indemnity policy to add explicit, affirmative cover for AI-related claims, reported as the first UK insurer to do so with a named, five-limb clause covering AI-generated advice, reliance on generative AI output, and failure of an AI system the insured designed (Insurance Business UK, 3 July 2025). That capability, clear wording naming AI scenarios instead of leaving them to guesswork, exists inside Hiscox's own product range. It has not reached the psychologist-facing policy it underwrites through Marsh Commercial, more than a year later.

Neither omission is unusual for a niche professional line. Insurance brokers writing about this market shift in mid-2026 note that where AI use has been disclosed and no specific exclusion applies, a policy would generally still be expected to respond, because silent cover is the default position until an insurer deliberately writes an exclusion in. The trouble is that "generally expected to respond" is a broker's read of how the market tends to behave, not a term printed in your policy. It is not the same thing as a confirmed answer from the insurer actually carrying your risk, and it is worth far less at the point a claim is being contested than a single email from your broker saying so in writing.

So what for you: the absence of an AI question on your policy isn't evidence that AI is a settled non-issue for your insurer. It's evidence that this specific product line hasn't been updated yet, while a sibling product from the same underwriter has. Don't read silence as an answer either way, confirm it.

The version of this that already exists, for doctors

UK doctors get a clearer answer from their own indemnity bodies than psychologists currently get from theirs. The Medical Defence Union, the medico-legal defence organisation for doctors and dentists, already tells its members what using an AI scribe changes about their medico-legal position: obtain and document the patient's consent to recording, review every AI-generated draft before it becomes part of the record, and apply your own clinical judgement to the content, because nothing an AI scribe produces counts as a clinical record until a clinician has reviewed and signed it. That framework, summarised in independent clinical coverage of NHS England's April 2025 guidance on AI-enabled ambient scribing, sits inside a different indemnity system to the commercial professional indemnity market that covers psychologists, and it doesn't extend to you. It's useful for a different reason: it shows what a mature, AI-specific answer from an indemnity provider actually looks like, and how far the commercial market covering independent psychology practice still has to go to catch up.

So what for you: don't wait for your own insurer to publish something equivalent. Build the same three habits, consent recorded, every draft reviewed before it's relied on, your own judgement applied, into your practice regardless of what your policy wording currently says, because they're the behaviour any insurer would want to see if a claim ever tested your use of AI.

The recommendation

The evidence points toward emailing your broker this week, not waiting for a renewal form that might not change in time. Write down which AI tool you use, what it does (transcription, drafting, summarising), who reviews its output before it's relied on, and where the data is processed. Send that as a disclosure, ask your broker or insurer to confirm in writing whether anything in your specific policy wording excludes or limits AI-related claims, and keep both the disclosure and the reply on file with a date on them. None of this requires a new policy or a higher premium to act on. It requires a paper trail that didn't exist an hour ago, and given that Professional Indemnity Insurance Brokers is already calling this one of 2026's biggest emerging underwriting themes, revisiting the question again at your next renewal, rather than assuming this week's answer holds indefinitely, is the safer default.

This sits alongside, not instead of, the clinical governance questions we've covered before: whether your client's consent to an AI tool is specific enough under HCPC's binding consent standard, and whether the ambient scribe itself has quietly become a regulated medical device under MHRA's guidance. Insurance is the piece that sits underneath both of those questions, ready to be tested the moment either one goes wrong. If you want a practice-specific read on where your own AI use and your own policy wording actually line up, our AI Opportunity and Growth Assessment covers governance and risk alongside clinical workflow. Or book a 20-minute call and we'll tell you honestly whether that's worth doing yet.

Related: HCPC's consent standard is binding. BPS's AI guidance isn't.  ·  Your AI scribe's status is set by marketing  ·  NHS's AI scribe guidance isn't binding

This article is for informational purposes only and does not constitute legal or insurance advice. It summarises publicly available UK professional indemnity market commentary (Professional Indemnity Insurance Brokers, July 2026; Miller Insurance, July 2026), the Insurance Act 2015's duty of fair presentation for non-consumer insurance contracts, and a direct check of the published coverage information for two named psychologist indemnity schemes on 5 September 2026. Policy wording changes between renewals and varies by insurer, broker and individual policy. Confirm your own position, including whether any AI-related exclusion applies, directly with your broker or insurer before relying on any interpretation above.