Chatbots, Hallucinations and 'Bad Actors': What the Fundraising Regulator's AI Guidance Actually Demands

The income pressure on UK charities is real. So is the temptation to automate away from it. Generative AI tools for appeal copy, AI-powered chatbots for donor FAQs, predictive models to rank major-gift prospects: all of them are arriving faster than sector governance has kept up. That gap just got a lot smaller.

As more charities explore using AI to support their fundraising, the Fundraising Regulator has published new guidance on the use of AI in charitable fundraising - the first time it has done so. The timing is not accidental. The Charity Digital Skills Report found that 76% of charities are using AI, up from 61% the previous year. The regulator decided that constitutes critical mass.

What the guidance actually says

The guidance takes a lifecycle approach. It applies whether a charity is exploring whether AI could support fundraising, preparing to implement AI tools, or already using AI - and is also relevant to charities that work with third parties who use AI, even if the charity itself does not.

Three specific use cases receive heightened scrutiny. Transparency is described as particularly important where AI-generated content could be mistaken by donors as real situations or people - such as sound, video and photographic images - where AI tools use personal data, or where chatbots are used for fundraising purposes. That covers a significant proportion of what development teams are actually deploying right now.

The regulator is blunt about systemic risks. Charities are warned about AI threats including misleading content, output influenced by hidden biases, and bad actors seeking to fundraise for criminal purposes. The regulator also warns against discriminatory or prejudicial content drawn from AI, as well as potentially inaccurate information, noting that AI could produce output that seems believable but is incorrect - known as hallucinations.

The Air Canada chatbot case is already being cited in sector legal circles as a cautionary analogy. AI can replicate racist and sexist biases and may seek to raise funds in an unethical manner, with documented examples of AI systems targeting vulnerable donors such as the recently bereaved and elderly. These carry the risk of legal action by donors, regulatory intervention, or significant reputational harm.

On oversight, the guidance is operationally specific. Charities must have a process in place for a human to check the accuracy, fairness and legality of any AI-generated content before using it for fundraising. Where there is no immediate human oversight - for example when using AI-powered virtual assistants like online chatbots or telephone voicebots - measures to monitor and audit AI uses must be in place.

On fundraising complaints or investigations, the regulator suggested charities keep a record of the checks they carry out and their decision-making in relation to AI-generated content, stating this will help justify actions and decisions if a complaint is received or fundraising is investigated.

CC20 closes the loop

The Fundraising Regulator's guidance does not sit in isolation. The Charity Commission updated its CC20 'Charity Fundraising: a guide to trustee duties' guidance on 3 February 2026. The updated guidance refers to the use of AI tools to write fundraising content and directs readers to the regulator's guidance if their charity plans to use AI for fundraising. Trustees are now explicitly on notice: AI in fundraising is a board-level governance matter, not a tool choice left to the digital team.

The guidance was made considerably more concise - from around 11,000 words down to 4,000 - and more accessible. The Commission framed this in terms of the workload context: the updates were designed to help trustees clearly understand their responsibilities, especially at a time when charities may need to work harder for donations whilst also seeing increased demand for their services.

The Fundraising Regulator's guidance highlights the importance of involving trustee boards in strategic decisions relating to exploring or adopting AI, and of building trustee understanding so that boards can adequately evaluate the range of AI opportunities and risks. The AI policy belongs on the board agenda, not buried in a comms team Notion doc.

The CRM gap

Here is where the regulatory signal meets a live operational problem. Most UK charity CRMs still lack meaningful AI features. The only UK-built platform currently offering AI across case notes, grant assessment, due diligence and impact reporting - with UK/EU-hosted data and GDPR-first design - is Plinth. Blackbaud's Raiser's Edge NXT has introduced some AI and machine learning features around donor analytics and fundraising optimisation, but its AI capabilities are limited compared to dedicated AI-first platforms, and data hosting defaults to the US, raising GDPR considerations for UK charities.

This matters for the regulator's transparency requirement. If your AI tooling sits inside a US-hosted CRM, sending personalised appeals or triggering chatbot responses off donor personal data, you have a simultaneous GDPR exposure and a Code of Fundraising Practice obligation to disclose. One of the most important considerations for charities is if AI can access personal data and how that will be managed securely.

Trademark filing data adds another dimension to the market picture. AIBD analysis of IPO (TMD) data shows Class 36 UK trademark filings - the class covering financial and fundraising services - reached just 835 in Q3 2026, down 65% on the prior period. That is a striking contraction in brand-protection activity at precisely the moment new AI-powered fundraising tools are entering the market. Whether it reflects a general slowdown in fintech and fundraising-tech venture activity, or simply timing, development directors should take note: many of the vendors pitching AI fundraising tools right now may not be building durable, protected brands.

What to do this week

Organisations planning to use AI are advised to test it first and develop and agree an AI policy, then publish it on their website to maintain trust and transparency. That policy should be kept regularly under review so it remains up to date and reflects all uses of AI for fundraising purposes.

If you already have a chatbot on your donation page, or a generative AI tool drafting appeal emails, audit it against the Fundraising Regulator's checklist before your next trustee meeting. Check whether it can access donor personal data. Confirm there is a human sign-off step before anything goes live. Check whether your data hosting is UK or EU-based.

The Fundraising Regulator has emphasised that existing fundraising standards continue to apply regardless of AI use, and charities should be open about their use of AI where it may be relevant to supporters. None of this is optional. The compliance trail the regulator now expects you to keep means ignorance is no longer a credible defence.