More than seven in 10 UK employees (71%) have used an unapproved artificial intelligence (AI) tool at work. These findings from the October 2025 Microsoft and Census-wide survey should give every charity CEO pause. Not because AI is inherently dangerous, but because what we don’t know can hurt us.
The philanthropic sector is embracing AI at an alarming rate. According to the Charity Digital Skills Report 2025, 76% of UK charities are currently using some form of AI. However, while in mid-2024, only 6% had developed an AI policy to govern their use of AI, it was encouraging to see that by 2025, around 50% of charities had developed a policy.
Still, the statistics reveal an unpleasant reality. Much of the adoption of AI in charities happens without oversight or guidelines, and often without anyone in leadership even knowing it was happening.
hidden workforce
Staff across industries are quietly incorporating AI into their daily work. They use it to draft emails, summarize reports, analyze data, and write funding applications. Most of them haven’t told anyone.
This silence is not surprising if you look at the numbers. According to a study by Microsoft and LinkedIn’s Work Trend Index, 52% of people using AI in the workplace do not want to allow AI to be used for critical tasks because they fear it will make them look replaceable or incompetent.
Even more shocking, according to research from The Access Group, 35% admit to secretly using AI for tasks they “should have done themselves.” The Joseph Rountree Foundation found that 73% of nonprofits have no AI guidelines at all.
In this vacuum, staff make their own rules. Some people think AI is okay because no one is saying otherwise. Some remain silent, fearing they will be frowned upon. Some people actively hide their use because they fear the repercussions of disclosure.
This is not a story about rogue employees or ethical failures. It’s about organizational design. If the charity does not provide clear explanations, staff will fill the gap with silence.
Why this is more important for charities
Every organization faces AI governance challenges, but charities operate in a uniquely sensitive area. The data that charities handle often falls into the GDPR’s ‘special categories’ – information about health, ethnic origin and religious beliefs. Protected records often contain sensitive data like this. When caseworkers paste client notes into AI tools to help draft reports, they may be sharing protected data with third-party processors without realizing the implications.
The regulatory environment is strengthening. In December 2025, the UK Fundraising Regulator updated its guidance on the use of AI in fundraising, making it clear that controllers are accountable for how their organizations use AI in this area. This is no longer a theoretical concern, but a governance obligation placed on boards.
But in reality, the use of AI is moving ahead of governance. One grant-writing consultant we spoke to explained that she received her client’s application and immediately realized what had happened. “That just stinks of ChatGPT. It’s clear that you literally just told me to write it and sent it to me… You’re going to send this to a panel of people now, and if I can see it, there’s a good chance they can see it too.” If staff are using AI without disclosing it, and external reviewers can quickly spot it, it creates a liability gap that management is clearly responsible for filling.
For charities, reputation is their operational infrastructure. The sector operates on trust from donors, beneficiaries, funders and the public. AI-related data breach and abuse incidents can cause lasting damage.
Funders are taking note
Major funders are beginning to take positions on the use of AI in application and program delivery. Approaches vary and charities need to be careful.
BBC Children in Need has indicated that it may ask about the use of AI during the grant review process. This is not necessarily a negative thing, they want to understand how the organization is working, but it does mean that charities need to be prepared to answer honestly.
UK Research and Innovation expects transparency around the use of AI in funding applications. This sets clear standards for charities involved in research partnerships to disclose their methodologies, including the use of AI tools.
The Esmée Fairbairn Foundation takes a different approach, stating that while it does not have a preference for the use of AI, it emphasizes the importance of being honest about how applications are developed.
The key message is that the positions of funders are becoming more diverse. Some are curious, some are cautious, but all seek candor. Charities that use AI without tracking or disclosing it can find themselves in awkward conversations. It’s better to get ahead of the curve now.
what to do now
The good news is that you don’t have to change your entire organization to address the use of covert AI. We need honest conversations and practical action.
First, audit your current AI usage. You cannot govern what you do not understand. Conduct anonymous staff surveys to find out what’s really going on across your organization. Ask not just whether people are using AI, but which tools they use for which tasks and what concerns they have.
For smaller charities, this may simply be a 15-minute team conversation rather than a formal investigation. The key is to bring to the surface what is happening, not create bureaucracy.
Put AI on the trustee agenda. This is not just an operational issue, but also a governance issue. Boards need to understand the AI landscape within their organizations and take responsibility for the associated risks and opportunities. One agenda item to start the conversation is a reasonable starting point.
Develop the right AI policy. The goal is not to ban AI completely or create bureaucratic hurdles. This is to clearly communicate to staff what is allowed, what requires approval, and what is prohibited.
A good policy eliminates the guesswork that increases confidentiality. For an overview of how different AI platforms handle sensitive data and how to protect each, see Hinchilla’s article on it.
Create a safe channel for staff to ask questions. When people fear judgment or consequences, they will remain silent. Make it easy and safe to ask for guidance. Consider appointing an AI lead or establishing a simple process for staff to check new tools before using them.
- Conduct anonymous staff surveys to understand AI usage
- Add AI governance to the agenda for your next board meeting
- Draft proportionate AI policies that enable rather than restrict.
- Create a clear, non-judgmental channel for staff questions
Bringing AI out of the shadows
None of this is about cornering people or turning back the clock on technology adoption. AI is already part of the fabric of philanthropy, but the question is whether it can be done transparently, securely, and in line with your values.
The staff who have been quietly utilizing AI are not the bad guys. They often find ways to work more effectively and perform well in resource-strapped organizations. The problem is not their actions. It’s the surrounding silence.
Charities have an opportunity to take the lead here. To demonstrate that responsible use of AI and effective use of AI are the same thing. Creating a culture where innovation happens in public, where risk is managed rather than ignored, and where staff feel confident to seek guidance.
The question is not whether your staff is using AI or not. It’s just a matter of knowing how.
