AI governance for a charity is the set of decisions, responsibilities, controls and evidence that helps trustees and staff understand where AI is used, why its use is appropriate and how harm, data misuse or loss of accountability will be prevented.
Questions the audit can help answer
- Which AI tools are staff and volunteers already using?
- Could sensitive beneficiary, supporter or employee information enter those tools?
- Who approves a new use and who can stop it?
- Which outputs require human review and what does meaningful review involve?
- Can trustees explain the organisation’s position using evidence rather than reassurance?
- Where could AI genuinely reduce avoidable administration without weakening relationships or judgement?
Good governance does not have to mean more bureaucracy
A small charity does not need to copy the systems of a multinational. It does need someone to take ownership, clear boundaries for staff and a sensible way to check higher-impact uses before they are adopted.
Common evidence
Useful evidence may include acceptable-use guidance, data-protection records, supplier terms, trustee papers, risk registers, procurement decisions, staff guidance, sample workflows and incident records. Missing documentation is not automatically a failure, but it affects what the organisation can demonstrate.
Where this shows up in practice
Service delivery
AI-assisted advice, triage, eligibility, safeguarding or casework can affect people directly and may require stronger oversight.
Fundraising and communications
Personalisation, drafting and supporter analysis raise questions about data, accuracy, consent and the organisation’s voice.
Internal administration
Lower-impact uses may still expose confidential material if boundaries and supplier terms are unclear.
Funding and reporting
Claims made to funders need reliable evidence and clear accountability, whether or not AI helped prepare them.
Trustee responsibility stays human
AI can support work, but it does not take over the responsibility of trustees or delegated decision-makers. The useful governance question is not simply “did we use AI?” but “was this use appropriate, controlled and capable of being challenged?”