Aipedia: Charities and trustees

AI chatbots giving beneficiaries advice

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

A charity giving advice through a chatbot or AI-drafted casework can be liable in negligence if the advice is wrong, and CC49 says it should consider professional indemnity insurance. Trustee indemnity alone is not designed for that risk.

Part of: AI in Charities and trustees

In short

Debt, benefits, housing and health charities are testing chatbots and AI-drafted casework to reach more people. CC49 warns that a charity giving professional services or advice may be liable if it is negligent and should consider insurance, and that claims-made policies may need funding for years after the advice, possibly beyond dissolution. The Charity Commission says AI is not reliable enough for accurate legal advice and urges extra care with vulnerable beneficiaries. A negligence claim from a beneficiary would usually fall to PI rather than trustee indemnity, subject to how the policy defines the charity's services.

How advice charities use AI

Demand for free advice outstrips adviser time, so charities are turning to AI to:

run website or messaging chatbots that answer questions about debts, benefit entitlement, tenancy rights or health conditions;
triage enquiries and route people to the right service;
draft letters, budgets and case summaries for advisers to review;
produce self-help guides tailored to a person's answers.

The people using these services are often in crisis, may have limited literacy or English, and are likely to act on what they are told.

The duties that apply

CC49 s.5.6: a charity that gives professional services or advice may be liable if it is negligent and should consider insurance.
The Commission's AI blog says AI is not reliable enough for accurate legal advice, human oversight should catch material errors, and charities working with vulnerable people, children or medical data should take extra care.
CC3 s.8 and s.9.3: trustees must act with reasonable care and skill and remain collectively responsible for the service, even when it is delivered by a bot.
Automated decisions: UK GDPR Articles 22A to 22D, in force since 5 February 2026 under DUAA 2025 s.80, set safeguards where a significant decision about a person is based solely on automated processing, such as automated beneficiary triage or eligibility screening.

Advice in some areas, such as debt, is also subject to its own regulatory regime; see the debt advice Aipedia guides.

What can go wrong

Illustrative scenario: a housing charity's chatbot tells a tenant that a notice from their landlord is invalid and they can ignore it. The bot has misread the type of notice. The tenant takes no action, misses a deadline and loses the chance to defend possession proceedings. They later say they relied on the charity's advice, given through its own website, with no adviser involved.

Other risks: a benefits bot that applies out-of-date rules, a triage tool that routes an urgent case to a waiting list, and AI-drafted letters an adviser signs without reading closely. A disclaimer saying the bot is "information only" may not help if it plainly gave advice. The scenario also shows how a single bot can repeat the same error to many people before anyone notices.

How PI usually responds

A beneficiary's negligence claim is aimed at the advice service, so it usually falls under professional indemnity rather than trustee indemnity. PI and trustee indemnity protect different people against different allegations. Whether a PI policy responds depends on whether it describes the charity's services broadly enough to include advice delivered by a chatbot, and on any AI or technology exclusions.

CC49 notes that many PI policies are claims-made, so a policy must be in force when the claim arrives. A charity that winds up an advice service, or closes altogether, may need run-off cover funded for years after the advice was given, possibly beyond dissolution. Trustee indemnity may still matter if a claim alleges the trustees failed to oversee the service. See whether PI covers AI mistakes.

Practical controls

Tell your insurer the bot exists and what it does before launch.
Limit the bot to information and signposting unless a qualified adviser checks its output.
Build in hand-off to a person for urgent or complex cases.
Keep logs of conversations so you can see what was said.
Test the bot against current rules and retest when rules change.
Record the risk assessment in your charity AI policy.

Safeguarding disclosures to a bot need their own route; see AI in safeguarding notes.

How Apex helps

Apex arranges PI, trustee indemnity and D&O for charities. A broker can check whether your PI describes your advice services in a way that includes chatbot and AI-assisted advice, and how run-off would work. See the AI in charities hub and how trustee indemnity treats AI claims.

Sources

Frequently asked

Is a charity liable for what its chatbot says?

It can be. CC49 says a charity giving professional services or advice may be liable if it is negligent. A chatbot run by the charity on its own website is likely to be treated as the charity's service, whatever the disclaimer says.

Does trustee indemnity cover a beneficiary advice claim?

Usually not on its own. Trustee indemnity covers trustees' personal liability for breach of trust or duty. A negligence claim about advice given to a beneficiary is normally a PI matter, subject to how the PI policy defines the charity's services.

What happens to cover if the advice service closes?

CC49 notes that many PI policies are claims-made, so cover must be in place when a claim is made, not when the advice was given. Trustees may need to fund run-off cover for years afterwards, possibly beyond the charity's dissolution.

Can AI give legal advice to beneficiaries?

The Charity Commission's AI blog says AI is not reliable enough for accurate legal advice. A charity using AI in legal or quasi-legal advice should keep a qualified person checking the output and be clear with beneficiaries about what the tool can and cannot do.

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.