“We’ve been told we can’t touch AI”
October 14, 2026 |
It’s not been a great few weeks for anyone trying to make the calm, sensible case for AI tools. Researchers have been quitting the big labs, including Anthropic, and going public with warnings that the technology they’re building could genuinely threaten humanity within a few years. That’s not a fringe view any more, it’s coming from the people closest to the work, and it’s genuinely alarming. And personally speaking, Trump saying ‘nothing to see here’ only raises that feeling of unease.
So it’s no wonder that charities, already careful custodians of some of the most sensitive data around, are getting more nervous rather than less. And here’s the awkward truth: that nervousness is likely to keep growing at the same time as the opportunity and the benefit does. Those two things are both true.
I’ve heard a version of this anxiety talking to clients over the past few months, including a large health charity saying, “we’ve been told we can’t touch AI.” It’s said with a shrug, and eye-roll, possibly a bit of relief, like they’ve been let off the hook.
Health charities are dealing with some of the most sensitive information any organisation handles: diagnoses, mental health, family circumstances, often shared by people at a difficult moment in their lives. And on balance, they tend to be pro the survival of humanity! So, the caution is absolutely fair enough. But there’s a big difference between AI concerns and taking data seriously and refusing to make a decision at all. Avoiding AI tools isn’t a sound policy, it’s the absence of one. And no policy doesn’t mean no risk; it just means the risk sits there unmanaged instead of properly governed.
Meanwhile, every month spent stuck on avoidance, on not grasping this nettle, is a month some other organisation is pragmatically rolling up sleeves and getting on with it: building better supporter relationships, faster insight, more efficient teams. And that gap doesn’t close itself.
We’ve been through this properly with Diabetes UK, building a set audience “Minds”, personas trained on their own segmentation and research, designed to bring the patient and supporter voice into everyday decisions across the organisation. It took months to get the governance right. It was meant to. The good news is that process is entirely repeatable, and that’s really the point I’m trying to make: get the groundwork done properly, and you’re set up to move forward with confidence. Sit on your hands, and you’re just deferring a decision you’ll have to make eventually anyway.
How Diabetes UK actually did it
To their enormous credit, Diabetes UK didn’t cut a single corner, which is exactly why the project has landed so well. They faced these challenges and concerns head on. Here’s roughly how it went:
They started with people, not paperwork. Before anything formal, sessions were run with key stakeholders to socialise what the Minds actually did or could do, and to get concerns and misconceptions out in the open early. The most important thing that came out of those conversations was a clear line between when an AI persona was useful, and when nothing should replace a real lived experience group. That distinction shaped everything that followed.
Then came a full Data Protection Impact Assessment, done properly rather than a box ticking exercise.
From there it went to their Information Governance Management Group, not once but twice. The first paper laid out the precedent, the data journey and the risks. Rather than trying to settle everything in one meeting, they agreed a second paper would dig into the areas that mattered most: could their data ever end up training AI for other clients (no, and that’s written into the contract), what does the ICO actually think about AI and personal data, and does any of this count as processing personal data in the first place (it doesn’t, the Minds are built from anonymised, aggregated insight, not individual records).
The proposal then went to the Executive team for review. What impressed me was that the conversation wasn’t focused solely on what was technically possible, but also on the broader ethical and governance considerations. The resulting decision was made thoughtfully, with clear ownership and accountability.
Finally, a contract that held all of this to account, and an Equity, Diversity and Inclusion assessment to close things out.
The bit that matters just as much: guardrails for real people using it
None of that governance means much if it falls apart the moment someone actually opens the tool. So alongside the formal sign off, Diabetes UK built simple, plain English rules for anyone testing the Minds. British English throughout. Never offer medical advice. Always point people to a GP or the NHS for anything clinical. No language that implies blame. And, most importantly, a very clear instruction that these tools are there to support lived experience, never replace it.
That last bit is the one I’d want every fundraising team to hear. This was never about swapping real people for a chatbot. It was about giving busy teams quicker, easier access to insight they already had, so that lived experience shapes more decisions, more often, not fewer.
Was it worth it?
Diabetes UK now has six rich personas built from years of segmentation, research and lived experience, plus an Insight Expert that can pull findings together across multiple studies. Feedback from early testing has been brilliant. As James Beeby, their Director of Engagement & Fundraising, put it: “This is epic… I have been blown away by the responses from the panel… a very big WOW!”
I’d argue that reaction exists because of careful governance, not despite it. It’s why the excitement feels earned rather than reckless.
A trojan horse, if you want one
Here’s what I’d say to any fundraising director sitting on an unresolved “no” to AI: a well scoped pilot is a brilliant excuse to answer the governance questions your organisation needs to answer sooner or later anyway. Consent, ICO alignment, where AI sits alongside lived experience, how much risk your board can stomach. Those conversations happen far more productively when they’re tied to something real, rather than left as an abstract debate that never quite gets resolved.
Do it once, properly, and you get more than a working project. You also get the muscle to move quicker on the next one, and in this fast-evolving and sometimes alarming world, that next one will come around the corner fast. The charities treating this as urgent now will spend the next few years with more confidence, building richer, more efficient, more human supporter relationships. The ones still waiting for permission will be explaining the delay to their trustees instead.
If any of this sounds familiar, and you’d like to hear more, let’s have a conversation.
This content was provided by 11 London
Company Details
Latest Content from 11 London
Your AI pilot works. That’s the problem.
What breaks when you scale it won’t be the model. It’ll be your data.
“We’ve been told we can’t touch AI”
It’s not been a great few weeks for anyone trying to make the calm, sensible case for AI tools. Researchers have been quitting the big labs, including Anthropic, and going…
The Power of Plurality
A thought piece on plurality of voice, and rewiring the advertising agency of the modern age
Stories of sickness: applying illness narratives to health advertising
By Dr Stephanie Goley Introduction As advertisers, we know that health and pharmaceutical campaigns hold enormous influence over how the public perceives illness, wellness, and treatment. But too often, our…
Advertising For Women’s Health: Breaking Taboos and Driving Change
Our strategist Stephanie Goley, MSc, PhD takes a fascinating look at how advertising is breaking down stigmas around areas of women’s health such as menstruation and menopause – and helping…
Rain Free Days: A Migraine App Shaped by Real Experiences
Rain Free Days® is a free migraine support app created by Teva, 11 London, and the migraine community. It helps users track symptoms, spot patterns, and access personalised resources like…
11 London and B Corp
Three years ago, 11 London became the first UK health agency to gain B-Corp accreditation. CEO and Co-Founder, Matt Hunt reports back on life in the B-List…
King’s College London and 11 London collaborate on health psychology
We’re delighted to announce that we are working in partnership with the Health Psychology department at King’s College London. Professor Rona Moss-Morris is working with the agency on a shared…
How we’re making AI a part of life at 11 London
This article talks about how AI is being used more in advertising, showing how it can change the way things are done from planning out strategies to coming up with…
Metahumans – the future of empathetic education?
Technology is moving fast, and here at 11 London we’re moving with it. Now, Metahumans (powered by Unreal Engine 5) are just one of many innovations launching in the digital…