Can charities use ChatGPT safely?

Yes, charities can use ChatGPT safely, as long as there are clear boundaries, a written policy and trained staff. The risk usually isn't the tool itself. It's informal, unpoliced use, where nobody has reviewed what staff are entering or what outputs they are acting on.

What data should never go into ChatGPT?

Never enter safeguarding-sensitive information, personal records, confidential donor data, medical records or anything covered by GDPR into a public AI tool. If the data would cause harm in a data breach, it should not be typed into a chatbot. This is the single most important rule for any charity adopting AI.

What can charities use ChatGPT for?

Safe, high-value uses include drafting fundraising copy, summarising public documents, generating first-draft reports, brainstorming campaign ideas, writing social media posts, translating public-facing content and formatting data. These tasks benefit from AI without exposing sensitive information.

What does a safe AI adoption look like?

A safe adoption follows four steps:

• Identify which tasks are safe to delegate to AI and which must never touch it.
• Write a policy that names the tools staff may use and the data they must not enter.
• Train staff on the policy and on practical prompting.
• Review the policy every six months as tools and risks change.

How much does it cost to get this right?

An AI Governance Pack from SAO Consultancy costs £750 and includes a tailored policy, risk register and staff guidance, delivered in under a week. A full AI Operations Audit costs £1,500 and maps every workflow against safeguarding, GDPR and regulatory constraints over two weeks. Both are fixed-fee with no hidden costs.