Overview

There’s a lot of pressure to get on board with AI right now. 76% of charities are apparently using it in some form, up from 61% last year. The headlines make it sound like a no-brainer - save time, raise more money, free up your team for the work that matters.

But here’s the thing: 59% of those charities still have concerns about what they’ve signed up for. And the ethical costs? They rarely get a mention.

This isn’t a guide to adopting AI faster. It's a framework for deciding whether you should adopt it at all - and for feeling confident whatever you decide.

The quick version

Reasons charities adopt AI:

Saves time on repetitive tasks like first drafts and meeting notes
Can help clarify or tighten up human-written content
Useful for summarising long documents or research
May support stretched teams with admin

Reasons to think twice:

Significant environmental impact from energy use
Built on human work often taken without consent or payment
Data handling is frequently opaque
Confidently produces errors, bias and inaccuracies
Accountability for mistakes falls on your charity, not the tool

The rest of this piece unpacks both sides - but with a focus on the questions that don’t get asked often enough.

The question nobody’s asking

Most AI guidance jumps straight to "which tools should we use?" But there’s a bigger question worth sitting with first: does AI actually fit with what your charity stands for?

Not whether it’s useful. Whether it’s right.

Charities exist to make things better. That purpose should shape every decision you make - including the technology you choose. Yet most of the advice out there treats ethics as an afterthought, something to tick off once you’ve already committed.

We think it belongs at the start of the conversation.

Why caution makes sense

The environmental cost

AI needs a staggering amount of power to run. Training a single large language model can produce as much carbon as five cars over their entire lifetimes. And every time someone asks ChatGPT a question, it uses far more energy than a normal search.

If you’re an environmental charity, an organisation that cares about climate - or has made any kind of environmental commitment - this matters. Saving a few hours of admin while quietly increasing your carbon footprint isn’t the straightforward win it’s made out to be.

Where the content comes from

Generative AI doesn’t come from thin air. It’s trained on enormous datasets of human work - writing, art, photography, research - often scraped without permission, without payment, and without credit.

There’s a hidden human cost too. The systems rely on content moderation and data labelling, work that’s frequently outsourced to low-paid workers in the global south, sometimes exposing them to harmful content with minimal support.

If fair pay matters to your charity, the supply chain behind AI deserves a closer look.

What happens to your data

AI tools process information in ways that aren’t always clear - sometimes even to the companies behind them. When you put data into a generative AI system, you might be feeding future training datasets, sending it to third-party servers, or losing track of it altogether.

The Charity Commission is clear that charities working with vulnerable people or sensitive information need to think carefully here. But honestly, most AI tools don’t offer the kind of transparency that lets you make a properly informed choice.

Getting things wrong

Here’s something that doesn’t get enough airtime: AI can confidently produce inaccurate, plagiarised, biased or copyright-infringing content without realising there’s a problem. It doesn’t understand the truth in the human sense of understanding, belief, or consciousness. It just predicts what words are likely to come next.

For charities, accuracy isn’t a nice-to-have. A mistake in a grant application, a safeguarding policy or a message to beneficiaries could cause real harm. And when AI gets it wrong, your organisation carries the blame - not the software company.

What gets lost

Charities work because people care about other people. That relationship - slow, messy, human - is what makes this sector valuable.

AI is sold as a way to automate the boring stuff so your team can focus on what matters. But it often creates new work instead: checking outputs, fixing errors, managing the tech, explaining decisions that a machine made but can’t then justify.

The time savings can be real of course - sometimes. But naturally the cost to human connection rarely shows up in the business case.

Five questions to help you decide

If you’re weighing up whether to use AI, these might help.

What problem are you actually trying to solve?

Be honest. "We want to be more efficient" isn’t specific enough. If you can’t name a concrete problem, AI probably isn’t the answer - and you might be adopting it just because it feels like you should.

Have you tried simpler solutions first?

Before reaching for new technology, consider whether the issue could be fixed with clearer processes, better priorities, or the right people in the right roles. Tech is often used to patch over problems that deserve proper attention.

Does this fit with your values?

Think about what your charity stands for. If you care about the environment, fair work, data rights or putting people first, AI adoption might create tensions. You’ll need to be able to justify those trade-offs - to your board, your beneficiaries, and yourselves.

Can your board actually oversee this?

Trustees are ultimately accountable for how their charity uses AI. That means they need to understand it well enough to ask the right questions. Half of charities are now developing an AI policy, up from 16% last year - but a policy without genuine understanding doesn’t count for much.

Who’s telling you to do this, and why?

Most AI advice comes from people selling AI products, AI consultancy, or content designed to get clicks. Their incentives aren’t the same as yours. Your job is to look after your charity’s resources, protect the people you serve, and stay true to your mission.

It’s okay to say no

Sometimes the right answer is: not for us.

Maybe the environmental impact doesn’t sit right. Maybe the labour practices behind AI clash with what you believe in. Maybe the data risks are too high for the people you work with. Maybe you just don’t have a problem that justifies the trade-offs. Or maybe the pressure is coming from outside your organisation, not from any genuine need within it.

None of that is failure. It’s thoughtful leadership.

The Fundraising Regulator doesn't mind either way. Neither does the Charity Commission. You’re allowed to say no. You’re allowed to wait. You’re allowed to put your values ahead of what’s trending.

The charities that’ll struggle most with AI aren’t the ones that take their time. They’re the ones that rush in without thinking, or feel they have to adopt because everyone else is, or realise too late that efficiency came at a price they weren’t prepared to pay.

There’s another way

You don’t have to join the race. There’s a quieter alternative: doing less, better, with people.

That means accepting some things take time. It means investing in your team rather than tools. It means being honest about what you can realistically do, rather than automating your way around the problem. It means protecting the slow, relational work that charities exist to do.

This isn’t anti-technology. It’s anti technology that conflicts with your purpose.

The sector doesn’t need more AI-generated content, AI-optimised fundraising or AI-assisted decisions. It needs organisations with the confidence to choose a different path.

Want to talk through AI, ethics, or your wider digital questions? We’re always up for a virtual coffee and an honest conversation.

Making the AI decision: A framework for deciding