A shift is taking place in many organizations away from the big headlines.

It's not in the conferences or official data, but in the details of daily work, where artificial intelligence is sometimes used before it is discussed and impacts before it is organized. This article serves as a professional reading in one of the most silent and sensitive spaces, drawing intellectually from an article published on Charity Digital by Robin Warren, which has been revised and rephrased to preserve the author's and original source's literary rights.

Our guest article, written by Robin Warren, founder of Hinchilla, explores how the unauthorized use of artificial intelligence has spread within charitable organizations, just as it has in the broader economy and the charitable sector itself.

More than seven out of ten workers in the UK, or 71%, have used unapproved artificial intelligence tools at work. This result, from research conducted by Microsoft and Censuswide in October 2025, is enough to make every CEO of a charity pause, not because artificial intelligence is inherently dangerous, but because what we don't know could harm us.

The charitable sector has rapidly adopted artificial intelligence, and according to the Charity Digital Skills Report 2025, 76% of charities in the UK are using artificial intelligence today in some form. However, by mid-2024, only 6% had developed a policy governing this use, though it is encouraging that around 50% of charities were working on developing policies in this regard by 2025.

Nevertheless, these statistics reveal an uncomfortable reality: a large part of the adoption of artificial intelligence within charities has occurred without oversight, without guidance, and often without leadership even knowing that it is happening.

The hidden workforce

Workers throughout the sector are quietly integrating artificial intelligence into their daily tasks, using it to draft emails, summarize reports, analyze data, and yes, even in writing funding requests. Most of them do not tell anyone about it.

This silence is not surprising when we look at the numbers, as research from Microsoft and the Work Trend Index from LinkedIn found that 52% of those using artificial intelligence at work are reluctant to admit it when it comes to important tasks, fearing it might make them seem replaceable or incompetent.

Even more striking, research from The Access Group found that 35% admit to secretly using artificial intelligence for tasks they were 'supposed to do themselves,' while the Joseph Rowntree Foundation found that 73% of nonprofits have no guidance on artificial intelligence whatsoever.

In this void, workers create their own rules, as some assume using artificial intelligence is acceptable as long as no one says otherwise, while others believe it might be frowned upon, so they remain silent, and a few deliberately hide their usage for fear of what might result from disclosing it.

This is not a story of rogue workers or ethical failures but a tale of organizational design because when charities do not provide clarity, workers fill this gap with silence.

Why does this matter more in charities?

Every organization faces challenges in governing artificial intelligence, but charities operate in a uniquely sensitive space, and the data they deal with often falls into what the General Data Protection Regulation (GDPR) classifies as 'special categories,' such as information related to health conditions, ethnic origin, or religious beliefs.

Moreover, safeguarding records often contain this type of sensitive data, and when a worker in social cases copies a client’s notes into an artificial intelligence tool to assist in drafting a report, they may be sharing protected data with a third-party data processor without realizing the implications.

It is worth noting that the regulatory environment is moving toward increased stringency.

In December 2025, the UK Fundraising Regulator updated its guidelines regarding the use of artificial intelligence in fundraising, clarifying that trustees are accountable for how their organizations use artificial intelligence in this domain and thus, this is no longer a theoretical concern but a governance obligation for your board.

However, what is happening in practice is that the use of artificial intelligence precedes governance, as one grant writing consultant described a situation where they received a request from a client and immediately realized what had happened, describing it as:

"It reeks of ChatGPT, and there's no doubt about it, as I literally asked them to write you the text and then they sent it to me... and now you're going to send this to a panel of important people, and if I can notice the artificial polish, then there’s a good chance they will too."

When workers use artificial intelligence without disclosure, while external reviewers can detect it immediately, a gap in accountability gradually forms, and this is a gap that trustees are now explicitly responsible for closing.

For charities, reputation is not just an added value; it becomes a fundamental operating structure. In short, this sector operates on trust; the trust of donors, the trust of beneficiaries, the trust of funders, and the trust of the public, and any data breach or incident of misuse related to artificial intelligence could cause long-lasting damage.

Funders are watching

Major funders have begun to form their stances on the use of artificial intelligence in applications and program implementation, and their approaches are differing, which charities need to pay attention to.

For example, BBC Children in Need indicated that they may ask about the use of artificial intelligence during the grant application assessment process, not necessarily as a negative thing, because they want to understand how organizations operate, but it does mean that charities need to be prepared to answer honestly.

Furthermore, UK Research and Innovation expects a degree of transparency regarding the use of artificial intelligence in funding applications, and for charities engaged in research partnerships, this sets a clear standard: disclose your methodology, including your use of artificial intelligence tools.

Meanwhile, the Esmée Fairbairn Foundation has taken a different approach, announcing that they do not favor any particular stance on the use of artificial intelligence, but emphasized the importance of honesty in clarifying how applications are prepared.

The key message here is that funder attitudes are beginning to diverge, with some becoming curious, some cautious, and everyone wanting transparency, and charities that have used artificial intelligence without tracking or disclosure may find themselves in awkward conversations, so it’s best to preempt that now.

What immediate action is needed?

Tackling the hidden use of artificial intelligence does not require a comprehensive organizational change but does require candid conversation and practical steps.

Start by auditing your current use of artificial intelligence, as you cannot govern what you do not understand. Instead, conduct an anonymous internal survey among staff to find out what is actually happening within your organization, and do not just ask if people are using artificial intelligence; ask about the tools they are using, the tasks they are using them for, and the concerns they have.

For smaller charities, this might simply be a 15-minute group chat instead of a formal survey, and the goal is to uncover what is happening, not to create a new bureaucracy.

Put artificial intelligence on the agenda for trustees because this is a governance issue, not just an operational one, and boards need to understand the AI landscape within the organization, and take ownership of the risks and opportunities associated with it. One item on the agenda can be a reasonable starting point for opening this discussion.

Develop a balanced policy for artificial intelligence because the goal is not to ban it entirely or to create bureaucratic obstacles; rather, the aim is to give staff clarity on what is acceptable, what needs approval, and what is off-limits.

Good policy, in turn, removes the guessing space that drives secrecy, and to learn how different AI platforms handle sensitive data and how each one secures it, you can refer to an article by Hinchilla on this.

Create safe channels for staff to ask questions; if people are afraid of judgment or consequences, they will remain silent, and make seeking guidance easy and safe. From here, you might consider appointing an AI officer, or establishing a simple mechanism that allows staff to check before using new tools.

Establish a clear and judgment-free channel for staff questions.

Draft a balanced artificial intelligence policy that empowers rather than restricts.

Add artificial intelligence governance to the agenda of your next trustee meeting.

Conduct an anonymous internal survey among staff to understand the use of artificial intelligence.

Bringing AI out of the shadows

None of this is about exposing people’s mistakes or winding back the clock on technology adoption.

Artificial intelligence has already become part of how charities operate, and the question is not whether it is happening, but whether it is happening transparently, safely, and in line with your values.

On the other hand, the staff members who have used artificial intelligence quietly are not villains; many of them are high performers seeking ways to work more efficiently within organizations that already struggle with resource constraints, and the problem is not in their behavior but in the silence that surrounds it.

Charities have an opportunity here to lead and to demonstrate that responsible use of artificial intelligence and effective use can be one and the same, and to build cultures where innovation happens openly, risks are managed rather than ignored, and staff feel confident when seeking guidance.

The question is: Do you know how they are using it?

The question is not whether your staff are using artificial intelligence.

The real answer about the entry of artificial intelligence into the nonprofit sector describes its entrance without a clear framework to regulate its presence and define accountability.

The organization that precedes this shift with policy, surrounds it with transparency, and ties it to governance does not just protect its data and reputation; it also safeguards the trust of its stakeholders in a sector that fundamentally relies on reliability, and the more important question will not be: Are you using artificial intelligence? But: Are we using it in a way that is professionally, ethically, and institutionally defensible?

This article has been translated with adaptations from charitydigital.org.uk