Professional indemnity insurance and AI use: questions to ask your insurer
By the ChamberOne team 5 minute read
Most firms adopt AI tools without telling their insurer. Nothing in the proposal form asked, so nothing was said. Then a claim arises that involves a draft produced with AI help, and the first conversation with the insurer is about what the firm knew, what it disclosed and what it had in place. It is a poor moment to learn the answers. The time to ask is before there is anything to claim for.
Key points
- Ask your broker in writing whether AI use should be disclosed, and keep the reply.
- Put specific questions to the insurer about exclusions, conditions, vendor errors and data incidents.
- Build a small file now: policy, tool list, training records, audit record and incident procedure.
- Notify early if something goes wrong. Check your policy for notification duties.
Why this is a conversation to have now
Professional indemnity policies are contracts, and their terms vary between insurers, between markets and between renewals. Insurers are still working out how to treat AI. Some ask direct questions on proposal forms. Some say little. Some may add wording over time. A firm that raises the topic early is in a stronger position than one that is surprised by it.
This article is general information, not advice on your policy. Your broker and the policy documents are the authority for your cover.
Disclosure at renewal
Proposal forms usually ask about changes in the firm's practice and systems, and about anything that might affect risk. Whether AI use falls within a particular question is a matter for your policy and broker. A cautious approach is to ask the broker directly, in writing, whether AI use should be disclosed and how, and to keep the reply.
Be accurate. A description such as we use an approved tool for summarising and first drafts, with lawyer review of everything before it goes out, is worth far more than a vague assurance, and it must be true. An overstated description is a risk in itself.
Questions to put to your insurer or broker
- Does our policy treat claims arising from work prepared with AI assistance any differently from other claims?
- Is there any exclusion, limitation or condition that refers to AI, automated tools or technology errors?
- Do you want to be told when we introduce a new AI tool, or only at renewal?
- What do you expect us to have in place, such as a policy, training or review procedures?
- Are there tools or uses that you regard as unacceptable, such as using public consumer chatbots with client data?
- If a claim involves AI output, what records will you ask us to produce?
- Does cover extend to errors by a vendor's tool that we relied on, and how does that interact with the vendor's own liability terms?
- Is there separate cover, or a cyber policy, that responds to a data incident involving an AI provider?
Where gaps tend to appear
Think about where AI-related loss might fall between policies. A missed deadline caused by a wrong date might be a professional negligence matter. A data leak caused by pasting client material into the wrong tool might be a cyber or data matter, or both. A fabricated citation in a court filing may bring costs orders and regulatory attention in addition to a client claim. Ask your broker to walk through a few such scenarios and say which policy would respond, and what would fall outside.
Do not assume that a vendor's contract will cover you. Many vendors limit their liability heavily. Read that clause when you buy, and tell your broker what it says.
Documentation insurers may want to see
Whatever the policy says, a firm that can show good practice is better placed than one that cannot. Build a small file now.
- 1
A written AI policy: approved tools, prohibited uses, supervision and review requirements.
- 2
A list of tools in use, with the provider, the purpose and who approved it.
- 3
Training records: who attended what, and when.
- 4
An audit record of AI use, showing who used which tool on which matter.
- 5
Examples of review in practice, such as checklists completed on real files.
- 6
An incident procedure: what staff do if they think they have made a mistake with AI.
Do not rely on a policy you have not shown anyone
A written policy that the staff have never seen is not a control. Insurers and regulators look for evidence that people knew the rules and followed them. Circulate the policy, have staff confirm they have read it and revisit it at least annually, because tools and risks move quickly.
- 1Circulate the policy to staff
- 2Have staff confirm they have read it
- 3Check that people knew the rules and followed them
- 4Revisit it at least annually
- 5Update it as tools and risks move
- Then back to step 1.
Handling an incident
Suppose a lawyer realises that a letter sent last week relied on an AI summary that misstated a date. The reflex is to quietly correct it. Policies generally impose duties to notify the insurer of circumstances that may give rise to a claim, and the timing can matter. Check your policy and contact your broker early. Tell the client promptly, as your professional duties require. Preserve the records. Speed and honesty usually reduce the damage far more than silence.
- Check your policy and contact your broker early
- Tell the client promptly, as your duties require
- Preserve the records
- Quietly correct the error
- Stay silent about it
Supervision is where claims are won or lost
When an insurer looks at an AI-related claim, a central question is likely to be whether the firm supervised the work as it would any other. A lawyer who checked the output, kept a note of what was checked and followed a written procedure is in a very different position from one who pasted a draft into an email. Make supervision visible. A simple checklist on the matter file, completed and dated, costs a minute and may matter a great deal later. The same goes for how junior staff are directed: say what tools they may use and who reviews their work.
What insurance does not do
Cover pays for losses after the event. It does not repair a client relationship, restore a reputation or remove a regulator's interest. The better protection is a process that prevents the error. Insurance is the safety net, and conversations with your insurer are most useful when they are part of a wider habit of asking what could go wrong and writing down what you have done about it.
This article is general information about the law, legal practice and technology, current when it was published. It is not legal advice. Laws and guidance change: check the sources linked above and take advice on your own circumstances from a lawyer qualified where you are.
Read next
Sri Lanka's Anti-Corruption Act 2023: a guide for practitioners
What the Anti-Corruption Act No. 9 of 2023 provides, how CIABOC exercises its powers, and what advisers must understand about offences, declarations, and a proposed 2026 amendment.
6 minute readClient confidentiality and AI tools: a checklist before you paste anything in
The questions to answer about retention, training use, access, hosting and sub-processors before client information goes into any AI tool.
5 minute readShadow AI in law firms: staff are already using it, so what next?
Why bans fail, how to find out what is being used, and how to offer an approved alternative with clear rules.
5 minute read