Calling all charities: Don’t let AI do your thinking

Artificial intelligence is transforming the charity sector.

From writing copy and analysing fundraising data to generating campaign ideas and improving productivity, AI is quickly becoming part of everyday life for fundraisers and marketers.

Used well, it’s one of the most valuable tools available.

Used badly, it can weaken decision-making, dilute creativity and undermine authenticity.

In this article, Richard Lee shares why charities should embrace AI as a powerful assistant, but never allow it to replace the experience, judgement and human understanding that successful fundraising depends on.

I think we’re approaching a critical point with AI in the third sector.

Some charities are using it to improve their thinking, and what a fantastic tool it is if used effectively. But there are others employing it to replace individual thinking and real-life experience and that’s a very dangerous move. Let me explain… 

We recently spent several days creating a proposal for an organisation.

It wasn’t something we’d thrown together over a couple of hours. It had been built on experience, conversations, campaign data, commercial modelling and more than a decade of working across both digital and face-to-face fundraising. 

Every recommendation had a reason behind it, even if that reason wasn’t written down in the document itself.

A little while later we discovered the proposal we submitted had been uploaded into ChatGPT with a simple instruction: “Critique this strategy.”

Now, before anyone thinks this is about to become an anti-AI article, let me reassure you that it isn’t. In fact, quite the opposite.

At Reach Fundraising we use AI every single day and, if I’m honest, I wouldn’t want to work without it now. It’s become another member of the team; one that can analyse thousands of rows of data in minutes, organise research, summarise lengthy reports, help us explore creative ideas and even build visual concepts before a designer spends time turning them into finished artwork. 

If I’m struggling with a headline or the structure of an article, it’ll happily throw twenty ideas at me in the time it takes to make a fresh brew.

Used properly, it’s one of the most useful business tools I’ve seen in my career.

The important words there though are ‘used properly’.

Like most people, I spent the first few weeks asking it ridiculous questions. 

I got it to write football chants, tell jokes and produce poems that thankfully will never see the light of day. Then, gradually, I started trusting it with more meaningful work, not because I wanted it to replace what I do, but because I realised it could help me do it better.

That’s still how I see AI today.

AI is a collaborator. It’s a sounding board. It’s a remarkably capable assistant.

What it isn’t, and should never be, is the person who makes the final decision. 

That’s where I think some charities, and plenty of businesses too, are beginning to lose their way.

According to McKinsey’s latest global survey, almost nine in ten organisations now use AI in at least one part of their business, yet only around a third have embedded it deeply enough to create meaningful value across the organisation. So according to the data almost everybody is using AI but far fewer have really worked out how to use it well.

I’m beginning to see that play out in the third sector.

Reports are uploaded into ChatGPT before anyone has discussed them. Strategies are judged by AI before the people in the room have challenged the ideas themselves. And proposals are critiqued by a machine that has no understanding of the conversations, compromises and commercial realities that shaped them in the first place.

Within seconds of upload the response comes back. 

It’s articulate, well-structured and written with such confidence that it’s very easy to assume it must also be right. 

That’s precisely where the danger lies, because confidence and correctness are two very different things.

The problem with AI isn’t that it sometimes gets things wrong. We’re all capable of that.

The problem is that it has no way of knowing what it doesn’t know.

It doesn’t know the board discussion that changed the brief halfway through the project. It doesn’t know about the campaign that failed last year and why nobody wants to repeat that shambles. It doesn’t know that a recommendation was deliberately excluded because of budget, regulation or internal politics. 

Unless somebody explains all of that, AI simply fills in the gaps as best it can.

Sometimes those assumptions are remarkably accurate but they can also miss the point completely.

The irony is that I see the opposite happen every time I work with AI.

Take this article and its staccato sentences. God, AI loves short, sharp, repetitive copy.

The first draft was perfectly respectable. The grammar was good, the structure made sense and most of the arguments were there. The problem was that it didn’t sound like me. The flow felt too mechanical, some of the paragraphs were too abrupt and every other sentence seemed desperate to become a quote for LinkedIn.

So I pushed back.

I explained the tone I wanted. I told it where the argument needed to breathe, flow and develop and where it was trying too hard to be clever. We went backwards and forwards, refining ideas, adding examples and challenging assumptions until the article gradually became something neither of us would have produced on our own.

That’s AI at its best. It doesn’t replace experience but it can enhance it.

One of the biggest misconceptions about AI is that it’s there to give you answers. 

In reality, it’s far more valuable when it helps you ask better questions – that’s when it becomes a tool with power. The best conversations I have with AI aren’t one prompt long. They’re twenty prompts long. I challenge it, it challenges me, we reject ideas, improve others and gradually work towards something stronger. 

In many ways, it’s no different from working with a colleague whose opinion you respect but don’t automatically accept.

AI works for inspiration not authenticity

The same applies to imagery.

Some of the AI image generation tools available today are extraordinary, and we’ve used them ourselves to explore campaign concepts, visualise ideas for clients and test creative directions before investing in photography or design. They save time, encourage creativity and often help clients understand an idea much more quickly than words ever could.

But charities should also recognise where the line is.

If AI is helping you visualise a campaign, that’s one thing. If it’s inventing beneficiaries, creating synthetic emotion or generating images that imply events which never happened, that’s something entirely different. Charities trade on trust and authenticity. Those aren’t marketing buzzwords; they’re the foundations on which supporter relationships are built, and they’re not worth compromising for the sake of convenience.

Perhaps my biggest concern isn’t that AI will replace fundraisers.

I honestly don’t believe it ever will. 

Fundraising has always been about people. It’s about understanding what motivates someone to give, building relationships over time and recognising that a spreadsheet can tell you what happened without ever explaining why it happened. Experience, empathy and judgement still matter enormously, and I suspect they always will.

What worries me far more is the possibility that some of us slowly stop exercising those qualities ourselves.

If every report is uploaded into AI before we’ve read it properly, every proposal is judged by ChatGPT before we’ve formed our own opinion, and every piece of copy is accepted because the machine says it’s good, then we’ve quietly outsourced the one thing our organisations actually employ us for.

Our judgement. Our experience. Those hours on the clock that make all the difference when pursuing remarkable results.

That’s the bit AI can’t replicate.

Not because it isn’t intelligent but because it isn’t living your organisation every day. It doesn’t know your supporters, your colleagues or your history. The wins and the losses. It only knows what you’ve told it and often that isn’t enough.

Interestingly, AI itself seems to understand this better than many of its users do. Ask it the same question in five different ways and you’ll often receive five different answers. Give it more context and its response usually improves. Challenge its assumptions and it’ll happily rethink its position.

In other words, it expects a conversation.

Perhaps we should too.

Charities are about people – never forget that

I genuinely believe AI is one of the most important technologies our sector has seen, and I think charities that ignore it risk being left behind – the opportunities are simply too great. 

It can save countless hours, uncover insights hidden in data, stimulate creativity and remove much of the repetitive work that stops talented people focusing on what really matters.

What it shouldn’t do is replace thinking.

So by all means use AI. Learn it, challenge it and make it part of your everyday work. Let it help you become faster, more creative and more productive than you’ve ever been before.

Just don’t make the mistake of outsourcing your judgement to it.

Why? Because the charities that thrive over the next decade won’t simply be the ones using artificial intelligence.

They’ll be the ones who combine artificial intelligence with human intelligence, curiosity and experience.

After all, AI doesn’t know your charity or your supporters. You do.