Why We Can Be Optimistic About Britain's AI Future

I am optimistic about Britain and AI. The talent is here, the organisations getting it right exist in numbers, and the answer, consolidate demand and diversify supply, is within reach.

I am optimistic about Britain and AI. That may surprise anyone who has followed this series for the last few weeks, in which I have made a sustained case that we are getting AI implementation wrong. Both things are true. This issue discusses why.

I have spent previous issues of this newsletter making a sustained case that Britain must make a strategic shift to consolidate demand and diversify supply. I stand by every word of it. The strategy gap is real. The institutional capacity problem is real. The pilot purgatory, the misdiagnosed people problem, the readiness assessments measuring the wrong things: all of it is real, documented, and in urgent need of attention.

And yet. I am also optimistic about Britain's advances in AI and the energy it brings to delivering AI that improves business and society. Let me tell you why.

The supply-side answer is not enough on its own

The announcement of the Sovereign AI Unit was a milestone. So was OpenAI walking away from Stargate UK the week before. Together they say something important: dependence on foreign AI infrastructure is now a named strategic risk in Westminster, and there is political will to do something about it. Good.

But supply-side investment, on its own, is the smaller half of the problem. Only about one in six UK businesses has adopted AI in core operations. Compute without capability is not a complete strategy.

The argument of Making AI Work for Britain is that the country needs to do both halves of the job: consolidate demand so that buyers, not vendors, set the terms, and diversify supply so that no single provider captures our future. This is the lesson GDS learned in 2012. It is the lesson we seem to have unlearned.

What I have seen in the last two years

While I have been researching and writing the book, I have visited organisations across the country that are doing this right. Not perfectly. Not at the scale the ambition requires. But right. A hospital trust in the north of England that redesigned its patient flow process with AI, slowly and carefully, with nurses and porters and consultants all part of the design team from the beginning, and that is now running at measurably better outcomes eighteen months in. A local authority in the Midlands that built an internal AI capability team rather than outsourcing it, trained them, retained them, and is now the organisation that other councils phone when they want to know how to start. A mid-sized manufacturing business in Yorkshire whose operations director told me, with some pride, that they had deliberately chosen not to move fast, and that the care they took in the first year meant their second year was transformation rather than fire-fighting.

These organisations do not make headlines. They are not featured in government showcases. They are just quietly getting on with it. And they exist in much larger numbers than the narrative of British AI failure would suggest.

That's why I believe Britain's problem is not a shortage of people who know how to do this. It is a shortage of systems that spread what they know. It is a solvable problem if we choose to solve it.

The structural reasons for optimism

Beyond the organisations I have visited, there are structural reasons to be optimistic that do not receive enough attention. Britain has, by any international measure, an extraordinary concentration of research talent in AI and related fields. The work coming out of British universities, not just Oxford and Cambridge, but many others such as Edinburgh, Manchester, UCL, Imperial, Bristol, Southampton, and my own University, Exeter, is genuinely world-class. And the Digital Policy Alliance, where I serve as Research Director, brings together exactly the kind of practitioner-policy-academic collaboration that the countries succeeding with AI have found indispensable.

We also have something that is easy to underestimate: a public sector that, for all its dysfunction, still contains people of real capability and extraordinary public service motivation who want to get this right. The NHS clinicians who are building AI tools in their spare time because they can see what is possible. The civil servants who are quietly fighting procurement rules that make no sense for AI contracts. The educators who are trying to use AI to give every child something closer to personalised learning, without waiting for a policy framework to tell them they are allowed to.

What needs to change, and what you can do

But my optimism is conditional, not unconditional. It depends on three things happening that are not currently happening at sufficient scale or speed.

We need to get serious about implementation infrastructure: not more strategy documents, but the practical support networks, the embedded capability teams, and the honest after-action reviews of what has worked and what has not, that allow learning to spread from the organisations getting it right to the ones still struggling. We need to stop treating workforce transition as a footnote to the opportunity narrative and start treating it as the central policy challenge it actually is.

And we need our public institutions to start rebuilding the technical capacity that was hollowed out over the last three decades: not because the private sector cannot contribute, but because sustainable AI adoption requires organisations that understand what they are buying, can hold suppliers to account, and can adapt when circumstances change.

None of this is beyond us. All of it requires political will, institutional courage, and the kind of patient, unglamorous, implementation-focused leadership that this series of articles has been about from the beginning.

So when I'm asked why I have spent so much time and effort into creating yet another book, I have a simple answer: Because it matters so much. That is what Making AI Work for Britain makes the case for. Making AI Work for Britain, with a foreword by Lord Ranger of Northwood, is published next week on 28th April and available to pre-order now: amazon.co.uk/Making-AI-Work-Britain

I hope you will read it, argue with it, share it with people who have the power to act on it, and tell me where I have got it right...and where I have it wrong.

It has been a privilege to think through these ideas with you over the past few weeks. Thank you for reading, for responding, and for the quality of challenge you have brought to this conversation. That, too, is a huge reason for optimism.

If this series has been useful, the best thing you can do now is share details of the book with someone who needs to read it. Not as a favour to me, but because the argument in it needs to reach the people making decisions about AI in British organisations.

You'll find more details at FutureOfAI.uk or contact me directly at alan@alanbrown.net.