AI for the national health service: a sense of direction?

By Nitin Shrotri, AI Champion Network Coordinator and Retired Consultant Urological Surgeon

July 2026

As the world gallops, naye charges forward in unpredictable and unorthodox ways in the world of AI, the matter appears to subsume and consume us all. The National Health Service (NHS) in the United Kingdom needs to keep moving with the times, and it’s not easy at all. In fact, to be brutally, honest, the NHS is tottering on its last legs, and needs a strong injection of many things some of which include money, enthusiasm, senior staff, morale, and finally desire, a strong sense of desire from the incumbent government of the day to ensure its survival. This is where AI tools are very likely to be of help, in creating shorter and faster pathways, but there are many issues that hold us back.

The first is democracy, and our sense of propriety. Don’t get worried! I am a total supporter of Democracy! My reason is because democracy is slow moving, full of red tape and decision-making is often at snail’s pace, of course rightfully so. But add the need for torrent-speed, fast moving AI into this situation, and by the time an AI tool gets approval, it is already last year’s fashion. This cannot be allowed to hold the UK back. The creation of projects like Sovereign AI will be helpful, hopefully nullify or alleviate these issues, but they will be under pressure to do so for many national causes, which are multiple, the NHS being only one of them. Attention and movement is therefore divided, and what we need is a fast track that helps us move in real quick time.

In this already treacly situation, if you add the small matter of “who’s carrying the can”, there will be immediate looking around everyone’s shoulder, to see who actually is! This is where leadership will be required, perhaps a joint sort of leadership to lessen the load of responsibility which is undeniably huge, and the NHS, MHRA and NICE must come together in creating some form of Traffic light system to help enable NHS Trusts and GP organisations to be able to know what systems to use, without having to pay for that expertise and legal wrangles individually. I see that MS Co-Pilot has now been rolled out for NHS use, which is almost certainly a good thing!

But that brings us nicely to the next point, the third point, that is the frontier AI labs. These labs are a small group of those huge multinational organisations that already rule over most of our lives already. They are building the underlying intelligence systems that most future workflows, platforms and products will very likely utilise. These frontier labs are not building the same thing. They are focused on different areas, each of which are special and specific. They are building advanced foundation models, multi model systems, safer, most shareable models, models understanding 3-D environments, autonomic coding systems or highly efficient models that can run in more practical enterprise settings. Some of these companies are XAI, Anthropic, Open AI, World labs, Liquid AI and Reka, to name a few. It is likely that these large systems will build AI tools that are used within the health service, and if so, just like the use of MS Co-Pilot, we will become slaves to that particular system in the health service. That is a slightly scary thought, perhaps almost a ‘given” as it may mean that our entire health service may change in unknown ways. Job losses are envisaged, but I am hoping that job losses will be minimised by the need for human supervision, and also the opportunity will be used to offer human sentiments like compassion to our patients in the health service. At present, professionals are short of that precious commodity, Time.

The next issue of concern is third-party internet access. Using AI tools will mean access to private and personal data and information available to 3rd parties so that AI tools embedded in the system can utilise it to help NHS organisations make decisions. But this means inevitably that third parties will have access to this data and however strong the governance and ethics of these large multinational corporations are, it is likely that money will talk because in this world money always talks. If we were to have AI tools that worked within the closed confines of an organisation, an NHS trust or a GP Surgery behind a local firewall, or even the entire NHS platform on the NHS app, it would feel safer, as we would not be at the mercy of another. Perhaps I am naive, and it is unlikely that the large AI players will not have eventual access to our data. However, much we try our best to protect it.

My reasoning to create a national body of NHS AI Champions was to provide a supportive peer group that would help offer fabric and structure to our medical fraternity in terms of governance, ethics, best practice and such. With GMC complaints already starting to come in against the medical fraternity from the use of AI, I would hope that we as the national health service choose a set of AI tools like we have done in initiating the use of MS Copilot from Microsoft. The previously mentioned traffic light system will ensure that the users of AI systems are assured of personal safety. The traffic light system will also, in addition, ensure that those that are given the ‘red’ light are culled, and will stop wasting energy and other resources offering us an ecologically sustainable AI model of creation. Those culled products could be offered some form of compensation by the government to ensure that inventors do not go out of pocket.

Finally, there needs to be an element of Public-Private partnership, without the state being fleeced, as is often the case. This will help us with financing new ideas and projects. Our people, good, clever and intelligent British inventors and innovators must be fully supported and their projects encouraged to the full.

I hope those who are involved, find this useful in some way. Thank you.