“AI is taking our jobs.” Is it though?

Before I get into this, it is worth remembering that AI isn't actually new.

Back in 1950, Alan Turing published Computing Machinery and Intelligence and proposed what he called the “imitation game”. The idea was essentially, if a person communicating through text couldn't reliably tell whether they were talking to another human or a machine, could we say that the machine was demonstrating intelligence?

More than 75 years later, that question feels rather less theoretical.

ChatGPT, Claude, Gemini and countless other AI tools can produce text that sounds remarkably human. They can write reports, generate images, summarise meetings, analyse information and produce social media posts in seconds. As Anne in BBC's Amandaland has shown us, they can apparently even become a form of friend and guide when she asked ChatGPT for advice on how to speak to her teenage son about sex (for those that didn’t watch, safe to say it didn’t exactly go to plan).

So, if machines are getting increasingly good at producing the things that people like me are paid to produce, should I be worried? Apparently so.

“Are you worried about losing work to AI?” an ex-colleague asked me recently.

Over six months into being a freelance charity and purpose-led communications consultant, perhaps the prospect of the work drying up should have sent me into a minor existential crisis. Yet it didn't (and believe me, many other things have).

My answer was a pretty resounding “no”.

That's partly because I'm confident about what I offer, but it's also because I've never really viewed AI as a replacement for communications expertise. I see it as a tool. It can be a very powerful one, certainly, but still a tool.

And, for complete transparency, it's a tool I use occasionally, usually when I'm feeling overwhelmed or missing having colleagues around to bounce ideas off (how sad is that? Maybe I am Anne).

I write my own content, but I'll sometimes use AI to critique something I've written, spot typo’s, question whether something is clear or summarise the key points of a hefty report I never seem to find the time to read properly. The AI tools I use regularly are image text generators (then edit them myself) and functions of Canva such as background remover (as you can see from the image with this blog).

Sometimes using AI saves me a huge amount of time. Other times I've disappeared down an AI rabbit hole trying to improve something I'd already written, only to find that an hour later my original thinking has been pulled apart, the personality has disappeared and I'm left wondering why I didn't just trust what I'd written in the first place. More than once I've scrapped the whole exercise and gone back to almost exactly where I started.

And I think that's part of learning to use AI well. Just because it can suggest a different way of doing something doesn't mean that way is better. Knowing when to ignore it, challenge it or close the tab altogether is just as important as knowing how to use it. I know for those that have perfected the art of feeding in content to their AI model this isn’t necessarily issue but the writing is a key part of the job I enjoy.

I don't hide the fact that I use it occasionally, but increasingly I can understand why some people might.

Because alongside the conversation about what AI can do, there is a perhaps more unkind one emerging about what it says about us if we use it. I think we need to be quite careful about where that takes us.

So, having resisted adding yet another opinion to the seemingly endless conversation about AI, I'm finally jumping on the bandwagon. I'm sure the internet has been waiting patiently.

AI can create it, but does it actually work?

Generative AI models learn patterns from huge amounts of existing information and use those patterns to generate new combinations and outputs.

I've seen people use AI to design their garden, organisations and businesses use it to create posters, my own local authority use it to design graphics for a deckchair trail and even a friend use it to create illustrations for a children's book.

These are all creative tasks and, in most cases, the person or organisation feeding the AI tool was initially pleased with the result. But I've also seen some of those outputs struggle to land with the people they were actually intended for (garden plan aside, obviously. I don't think the plants had particularly strong feelings).

And that's where I think there is an important distinction between generating something novel and having an original perspective.

For communications, that matters enormously because good communications doesn't start with “what shall we post?” It starts with understanding who we're trying to reach, what matters to them, what might stop them engaging, what we want them to think, feel or do and what is happening around them that might affect how our message lands.

Only then comes the content; AI can help create that content, but somebody still needs to bring the understanding, context and judgement behind it.

AI doesn't know when it's wrong

One of the better-known problems with generative AI is that it can confidently produce information that simply isn't true. Humans get things wrong too, obviously. I've yet to meet the communications professional with a 100% accuracy record. The difference is that responsibility still sits with us.

Whatever tools I use along the way, I'm still responsible for what I hand over to a client. If I don't know whether something is correct, I check it. If the source looks questionable, I investigate further. If I'm communicating something sensitive, I think about the consequences of getting it wrong.

“ChatGPT said so” isn't much of a defence.

This is particularly important for VCSE organisations, where AI might be used around health information, safeguarding, funding, policy, sensitive data or people's lived experiences.

The Charity Commission has warned that generative AI can produce inaccurate, biased, plagiarised or copyright-infringing material while presenting it confidently, and its guidance stresses the importance of checking AI-generated content and maintaining appropriate human oversight.

Interestingly, the Commission has also reported seeing AI-generated charity registration applications that are too generic to properly reflect what an organisation actually does.

And these aren't the only concerns. There are important questions about the environmental impact of the energy and water required to develop and run AI systems too (that is part of a much bigger conversation).

So there are very good reasons to be cautious about AI, but caution and condemnation aren't the same thing.

When did using AI become something to be embarrassed about?

There is now so much generic AI-generated content online that we've developed a term for it: AI slop.

LinkedIn itself has used the term. In June 2026, the platform announced changes designed to tackle low-quality and inauthentic content, including what it described as content that “many call AI slop”. Its explanation was interesting: AI can help people with their writing, but what people share should still reflect their own voice, perspective and expertise.

For somebody working in communications, I find that fascinating.

We've developed technology capable of mimicking human communication at enormous scale, and one of the challenges facing the platforms distributing it is now how to find the actual humans amongst all the content. It's a crazy world, isn't it?

But I think there's another side to this conversation.

When “that's obviously AI” becomes an insult, we risk treating using AI badly and using AI at all as though they're the same thing.

They're not and that's where I think things become particularly interesting for the VCSE sector.

Demonising AI won't stop people using it

People are already using generative AI at work. They're using it to draft emails, summarise meetings, develop funding applications, analyse feedback, write reports, translate information, record minutes and get started when they're staring at the dreaded blank page.

Some of that use will be sensible and some of it probably won't be.

But if our response whenever we discover that someone has used AI is judgement, ridicule or an assumption that they've somehow cheated, I don't think people will suddenly stop using it.

I think they'll stop telling us. Surely that's more of a concern?

I've seen organisations starting to grapple with this in very different ways. Some have developed AI policies without really involving the staff who will actually be expected to follow them. Others haven't yet invested the time and energy in developing a shared approach at all.

Both potentially leave quite a gap between policy, practice and what leaders think is happening.

If people across an organisation are making their own decisions about which tools to use, what information to put into them and what is acceptable, while also worrying that colleagues might judge them for admitting they're using AI, that use doesn't disappear. It just gets pushed further under the dusty rug. Surely that's an organisational nightmare?

How can you safeguard your organisation if you don't have a proper grasp of the tools people are using and how they're using them? Would you know if a member of your team uploaded personal or sensitive information into a generative AI tool? More importantly, would they feel able to tell you if they realised afterwards that perhaps they shouldn't have?

For VCSE organisations in particular, where trust and accountability matter and teams may be handling sensitive information or communicating about people's lived experiences, those aren't insignificant questions.

This is why transparency matters.

It gives organisations the opportunity to ask sensible questions before something goes wrong. What are people using AI for? What information shouldn't be shared? What needs checking? When is human oversight essential? Where could bias cause harm? When is AI genuinely saving time and when are we using it simply because it's there?

I'd much rather someone said, “I used AI to help me structure this, then I checked it, changed it and made sure it reflected what we actually wanted to say,” than quietly copied and pasted something because they thought admitting to using ChatGPT would make them look lazy or less capable.

And if somebody does get it wrong, I'd want them to feel able to say so. If information has been shared that shouldn't have been, or inaccurate AI-generated content has made its way into something public, knowing about it means the organisation can actually do something about it.

A policy absolutely has a role to play, but a policy on its own isn't going to create responsible AI use. People need to understand it, have some ownership of it and, crucially, feel able to have honest conversations about what they're actually doing.

We can't manage something properly if nobody feels comfortable admitting that it's happening.

Perhaps one of the first questions for VCSE leaders should be: “Do we actually know how AI is being used in our organisation?”

And what happens to our own skills?

Being open about AI use doesn't mean embracing every possible use of it.

One thing I think we need to talk about more is what happens to our own skills and confidence when we routinely hand the thinking over to AI.

Writing, analysing, questioning, structuring an argument and coming up with ideas are things we get better at by doing them. They're also, for me at least, some of the most interesting parts of the work. The thinking, questioning, making connections and eventually landing on something that feels right is part of the process. If we hand all of that over to AI, we're not just potentially losing skills, we're taking some of the interesting bit out of the work too.

That doesn't mean using AI to tidy up your meeting notes is slowly turning your brain to mush. But I do think there is a danger in moving from “AI can help me do this” to “I don't think I can do this without AI”, until it becomes a comfort blanket we no longer feel confident working without.

If every blank page immediately gets handed to an AI tool, when do we practise finding our own starting point? If every sentence gets rewritten before we've decided whether we actually like it, do we gradually become less certain of our own voice? If AI generates the questions, structure, ideas and words, are we still using it as a tool or have we started outsourcing the thinking?

Again, those aren't reasons to demonise AI, they're reasons to talk more openly about how we're using it.

If AI saves us time, what are we saving it for?

For organisations where time, capacity and money are often in short supply, dismissing AI altogether doesn't make much sense to me either.

Used well, it can take some of the weight out of resource-heavy tasks, help people make sense of large amounts of information and give stretched teams a useful starting point. The Charity Commission itself recognises that, while AI brings risks, it also has the potential to free charities from resource-intensive tasks and create more time for higher-priority work.

And surely that's the opportunity?

Not to use AI to replace the thinking, judgement, creativity or relationships that make our work valuable, but to use it where it genuinely gives us more capacity for those things.

Although I think there's another conversation hiding in the idea of AI “saving us time”.

If something that previously took two hours now takes 30 minutes, do we have to immediately find another 90 minutes of work to fill the gap?

For organisations where people are already stretched, I'd be wary of AI simply allowing us to normalise an expectation that everyone should produce more. If every efficiency immediately gets swallowed up by another task, another meeting or another item on the to-do list, I'm not sure we've made working life much better.

Perhaps some of the time AI gives us back could actually be given back to people. More space to think properly, have a conversation with a colleague, develop an idea, build relationships or simply work at a pace that doesn't leave everyone permanently overwhelmed.

Efficiency shouldn't automatically mean more output from the same person.

And perhaps that's part of using AI responsibly too. It's not just about what we can get it to do, but being thoughtful about what we want to gain from using it in the first place.

There is a huge difference between using AI to help make sense of 500 responses to a consultation and asking it to tell you what your organisation's position should be. There's a difference between using it to sense-check something you've written and publishing its first response untouched.

The tool might be the same. The judgement around how we use it is what matters.

Which brings me back to where I started: we need to be able to talk honestly about how we're using it.

So, am I worried about AI taking my job?

No.

My clients don't just pay me to put words in the right order. They pay me to understand what they're trying to achieve, ask questions, spot gaps, challenge assumptions, understand different audiences, make connections and help them work out what they actually need to communicate in the first place.

AI doesn't replace that for me. It can sit around my work, helping me sense-check something I've written, find typo’s I've missed, but the thinking, judgement and final decisions are mine.

What concerns me much more is creating a culture where people feel they have to hide the fact they're using AI at all. Because if we demonise its use, I don't think we'll get rid of it. We'll just make people less likely to admit they're using it, and for organisations trying to manage accuracy, data, safeguarding, skills and trust, that lack of transparency could be far more problematic than the use of AI itself.

Perhaps the biggest risk isn't that people in our organisations are using AI. It's that they're already using it and we haven't created a culture where they feel able to tell us how.

So yes, I use AI occasionally. I think carefully about when I use it, what I use it for and when I'm better off without it and for me right now that is in the writing.

I'd much rather we had more honest conversations about AI than pretended none of us are using it, while making harsh judgements about those who openly admit they are.

The world feels divided enough without us finding another reason to judge each other. So perhaps we could replace some of that judgement with curiosity, both about how other people are using AI and how we're using it ourselves.

If you use AI regularly, maybe it's worth checking in occasionally. Is it a tool that's supporting your work, or has it gradually become a safety blanket you don't quite trust yourself to work without?

If it's the latter, perhaps that's worth paying attention to. Not because using AI is somehow cheating, but because your own skills, judgement and confidence are worth holding onto.

References

Turing, A. M. (1950), Computing Machinery and Intelligence, Mind, 59(236), 433–460.

Charity Commission for England and Wales (2024), Charities and artificial intelligence, Charity Commission blog.

Charity Commission for England and Wales (2025), An evolving charity sector, Charity Commission blog.

LinkedIn (2026), Keeping conversations real on LinkedIn.