I am a geriatrician, which means I spend my days with older people — usually polite, usually grateful — and a surprising amount of my work involves taking things away: the medications and interventions that no longer benefit the patient. It is good, fulfilling work. I would like to think my patients are fond of me. I am quite certain of one thing they would not be fond of: a mechanised robot careering toward them at speed to stop them taking an extra sachet of laxatives. Absurd as it sounds, that is genuinely my first thought whenever I am told that AI and robots will revolutionise a health service still struggling to move from paper to electronic notes.

What unsettles me is not the technology. It is the ideology riding on its back. When I am cornered by a techno-evangelist, at a conference, or in the dingy corridors of LinkedIn, what strikes me is how quickly a technical pitch curdles into something closer to faith.

At a recent conference, an industry leader offered a line I found genuinely disingenuous: that if you took the money going into AI and spent it instead on doubling the number of doctors and nurses, there still would not be enough to run hospitals properly. He went on: “Healthcare AI superintelligence will be better than anything humans can offer”. I do not doubt the technology’s potential. But when a member of the audience asked, reasonably, whether a system that clever would accept being sued when it harmed a patient, you could watch the speaker flinch. The cleverness, it seems, stops entirely at the threshold of accountability.

Now, let me take you to the real world. Hospitals. Patients on trolleys, crowded corridors, bleeps going off, staff running ragged and understaffed. Integrated care boards have been ordered to halve their running costs; we were promised the frontline would be spared, and yet the middle-grade doctors who hold the place together in many district general hospitals are quietly not having their contracts renewed. If the leaks about the workforce plan are right, nurses are not safe either.

Do not worry, people of Britain, AI will fix all of these structural problems; every one of them baked into a system running in a permanent state of crisis. It will, we are told, optimise workflows. Do my paperwork. Monitor observations and alert me when they drift out of range — something automated systems already do. Help me reach the right diagnosis, even though the genuinely complex diagnostic puzzle is, in practice, the exception and not the rule. And by doing all of this, it will somehow save the NHS.

This is not the grumbling of a lone sceptic. When Medscape surveyed UK doctors in 2024, around half said they would not use AI to treat patients, and more than 8 in 10 wanted government or professional oversight of its use. The medical profession is not frightened of the future. It is asking a simple question: who is accountable when the future goes wrong?

Here is where faith meets the balance sheet. Our regulator, the MHRA, is a genuinely excellent institution — from covid to vaping strategy, it has delivered, with its world-class expertise. Yet its entire annual budget would barely cover the production of a single Hollywood blockbuster (if the Guardians of the Galaxy can be called that), and it is being asked to hold the line against companies whose revenues rival the GDP of nation-states. That is not a fair fight, and everyone in it knows so.

Which brings us to the question that only a few people at these conferences care to dwell on. Do we, the British public, trust our American cousins so completely that we will hand over all of our medical data to their companies and simply take it on good faith that when people are ill, anxious and frightened, a healthcare AI will never slip in an advertisement for a particular drug or service? Imagine. That would be a captive market of extraordinary value: the sick, the worried and the dying, profiled and marketed to, at their most vulnerable. A goldmine, which we are asked to believe will go politely untouched. Set aside, for the moment, the fragile civilian power grids, dwindling fresh water supplies, and shaky semiconductor supply chains that the technology all depends on. The risks of AI misalignment (where it refuses to listen to its human supervisors) are another can of worms. The simpler problem is right in front of us — a healthcare system that will happily host any AI pilot offered to it, yet rarely has the money to deploy the slightly less exciting programmes (the ones that actually work).

For all my heresy, I am a believer. I believe in what artificial intelligence might one day do, and given the way the world is ageing, I think some marriage of AI and robotics will eventually be essential to keep healthcare affordable and, perhaps, universal.

But there is a line.

We cannot allow our patients, our hospitals, and our institutions to become products to be consumed by the technology or by the conglomerates that sell it. What we need first is an honest reckoning with what these systems can actually do today — in hospitals where you would struggle to find fully digitalised notes.

So, respect the technology but beware the techno-evangelists. They do not even need to be right. They only need to convince the people who hold the budgets, and when they do, the pattern will be the familiar one: privatised profits and socialised risks. Let us not, this time, give in to it.

Author

Shiva Pandey

Dr Shiva Pandey, MRCP, is a registrar in geriatric medicine and general internal medicine in the NHS, and a student on the Executive MSc in Health Economics, Policy and Management at the London School of Economics.

Declaration of Interests
I have previously undertaken strategy consultancy for the MHRA and sit on the steering committee of the CRUK E-Cigarette Research Forum.