Professional indemnity

AI risk factors in professional indemnity insurance

Reviewed by the Apex broking team · Last reviewed 2026-10-19

AI risk factors in professional indemnity are the ways a firm's use of AI tools raises the chance or size of a negligence, confidentiality or IP claim. They matter because underwriters now ask about AI use, and the answers shape terms, exclusions and whether the policy responds.

Part of: Professional indemnity insurance

In short

Using AI does not change the legal test: a professional must still exercise reasonable care and skill, and courts and regulators say the professional stays accountable for AI output. What changes is the risk profile. AI can produce plausible but wrong content, move client data to third parties, reproduce protected material and spread one error across many files. Underwriters commonly ask what tools you use, for what work, who checks the output and what your contracts promise. Most PI wordings were not written with AI in mind, so check how yours treats AI-assisted work alongside your cyber policy.

How AI changes professional liability risk

The duty does not move because a tool did the first draft. Where a business supplies a service to another business, section 13 of the Supply of Goods and Services Act 1982 implies a term that the supplier will carry out the service with reasonable care and skill. For consumer clients, section 49 of the Consumer Rights Act 2015 does the same job. The question in a claim is whether a reasonably competent member of your profession would have relied on that output without checking it. See our note on section 13 and PI.

The courts have been direct. In Ayinde v London Borough of Haringey [2025] EWHC 1383 (Admin), the Divisional Court said freely available generative AI tools are not capable of conducting reliable legal research, and that lawyers who use them have a professional duty to check accuracy against authoritative sources. The principle reads across to other professions.

AI changes four things underwriters care about:

Frequency: more output, produced faster, with less review time per item.
Severity and aggregation: a flawed template, prompt or model used across many clients can create a series of related claims from one root cause.
New heads of loss: data protection, confidentiality and IP claims that sit at the edge of PI and cyber cover.
Contract terms: clients increasingly write AI clauses into engagement terms, which can widen your liability.

Where liability arises

Negligent advice and hallucinated content

Generative tools can produce confident, plausible and wrong answers: invented case citations, misquoted standards, wrong tax rates or calculations. The client's claim is against the firm, not the tool. Check the AI supplier's terms too: if they cap or exclude the provider's liability, any recovery from the vendor may be limited. Our page on AI-generated advice and negligence covers the negligence analysis.

Confidentiality and data protection

Pasting client material into an external tool may breach confidentiality, privilege or a non-disclosure clause. Where personal data is involved, UK GDPR applies. The ICO's guidance on AI and data protection covers accountability, impact assessments, fairness, transparency and accuracy; the ICO says it is under review following the Data (Use and Access) Act 2025.

Automated decisions are a specific point. The old Article 22 of the UK GDPR has been replaced by Articles 22A to 22D, fully in force from 5 February 2026. Under Article 22C, where a significant decision is taken solely by automated processing, the controller must provide safeguards, including information about the decision, a way to make representations, human intervention and a route to contest it. A firm using AI to screen, score or decide about individuals needs to show meaningful human involvement or have those safeguards in place. Read more on GDPR and data breach cover within PI.

Copyright and IP infringement

Copyright is infringed by copying the whole or a substantial part of a work, directly or indirectly, without licence (Copyright, Designs and Patents Act 1988, s.16). AI-generated text, images or code can reproduce protected material, and firms that deliver it to clients may face infringement claims or indemnity demands. See AI-generated content and IP infringement.

What professional bodies expect

Solicitors: the SRA's Risk Outlook report on AI in the legal market (November 2023) warns about hallucination, says firms must protect confidentiality and privilege, and says firms remain responsible and accountable for AI outputs. It also advises firms to engage with their insurers about AI use.
Surveyors: the RICS professional standard Responsible use of artificial intelligence in surveying practice has applied to members and regulated firms since 9 March 2026. It covers governance and risk management, professional judgement and oversight, and transparency with clients, including risk registers, responsible use policies and procurement due diligence.
Accountants: ICAEW publishes a Generative AI guide for members, which includes governance and ethics.

Breaching professional-body requirements is not automatically negligence, but it is strong evidence of the standard expected, and underwriters commonly ask whether you follow it.

Does PI respond? Policy issues

Whether a particular AI-linked loss is covered is a separate question, covered in does professional indemnity cover AI mistakes? At portfolio level, these are the issues underwriters and brokers commonly work through, all subject to the policy terms.

Issue	Why it matters	What to check
Silent AI	Most wordings neither include nor exclude AI expressly, leaving room for dispute	Whether cover is triggered by your professional services however produced
Technology or AI exclusions	Some wordings exclude losses from software, data or automated systems	Exact wording; see AI exclusions and clauses
Cyber overlap	Data leaks via AI tools may fall under cyber, PI, both or neither	Which policy leads; see cyber insurance and AI risks
Contractual liability	PI commonly excludes liability you accept by contract beyond what you would owe at law	AI warranties, indemnities and uncapped clauses in client terms
Aggregation	One faulty process can produce many claims	How related claims are grouped against the limit

The silent AI issue mirrors the earlier move on silent cyber, where the market moved towards wordings that say expressly whether cyber is covered. For contract terms, see contractual liability in PI.

What underwriters commonly ask

Proposal forms and renewal questions vary, but underwriters commonly ask:

Which AI tools are used, and are they enterprise products or free public tools?
Which work types involve AI: research, drafting, calculations, client-facing chat, decisions about individuals?
Is there a written AI use policy, and is staff use monitored?
Who reviews AI output before it reaches a client, and how is that recorded?
Is client or personal data entered into AI tools, and on what supplier terms?
Do engagement letters tell clients about AI use and limit liability?
Have you given contractual warranties about AI, or accepted client AI clauses?
Do you sell AI-based products or advice, rather than just using AI internally?
Have there been any AI-related complaints, errors or near misses?

Answers form part of your duty of fair presentation, so describe AI use accurately rather than minimising it. Good records help; see what to document about AI use.

Common misunderstandings

"The software made the mistake, so it isn't our liability." The client relied on you. Regulators and the courts treat the professional as accountable for the output.
"PI covers anything we do." PI covers liability from your professional services, subject to exclusions. Data breaches, regulatory fines and contractual promises can fall outside it.
"Cyber will pick up anything involving AI." Cyber usually focuses on data, systems and breach costs, not negligent advice. See where cyber and PI cover begin and end.
"We don't use AI." Many everyday platforms now include AI features. Underwriters commonly expect firms to know where it is in their workflow.

What to check

Map where AI is used, by whom and for what work.
Adopt a written policy covering approved tools, data inputs and review steps, and follow any professional-body standard that applies.
Record human review of AI output on client files.
Review supplier terms on data use, training and liability.
Check engagement letters and client contracts for AI clauses, warranties and indemnities.
Read your PI wording for technology, data and contractual liability exclusions, and line it up with your cyber policy.
Disclose AI use accurately at renewal.

Sources

Frequently asked

How does AI affect professional indemnity insurance?

AI does not change the legal duty to use reasonable care and skill, but it changes the risk profile underwriters assess. It can increase error frequency, create aggregated claims from one faulty process, and add confidentiality, data protection and IP exposures. Insurers commonly ask about AI use at proposal and renewal, and terms may depend on your controls.

What do insurers ask about AI use?

Questions commonly cover which tools you use, whether they are enterprise or public products, what work they support, who checks the output, whether client or personal data goes into them, whether you have a written policy, and whether contracts contain AI warranties. Answer accurately, because your answers form part of your duty of fair presentation.

Is a professional liable if AI gets it wrong?

Usually, yes. The client relied on the professional, and the test is whether a reasonably competent professional would have used and checked the output as you did. The High Court in Ayinde (2025) said lawyers using AI have a professional duty to check accuracy against authoritative sources, and the SRA and RICS say the professional remains accountable.

What is silent AI in insurance?

Silent AI describes policies that neither expressly cover nor expressly exclude losses involving artificial intelligence. It leaves room for dispute about whether a claim falls within cover. It mirrors silent cyber, where the market moved towards wordings that say expressly whether cyber is covered. Check how your PI and cyber wordings treat AI-related losses.

Do surveyors have rules on using AI?

Yes. The RICS professional standard on responsible use of artificial intelligence in surveying practice has applied to RICS members and regulated firms since 9 March 2026. It covers governance and risk management, professional judgement and oversight, and transparency with clients, including risk registers, responsible use policies and due diligence when buying AI systems.

Need professional indemnity cover?

Start the proposal online, or upload last year's proposal on the first page. Or call 0117 325 0027.

Start your proposal 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, not advice on your individual circumstances, and it does not guarantee that cover will be available or on what terms.