The flaw in Wes Streeting's AI NHS app plan

Spectator

5 hours ago

Speaking at Blackpool Football Club earlier this week, Wes Streeting announced his latest bid to modernise the NHS: bold new additions to the NHS app. Artificial intelligence would be used to empower people, turning them into experts on their own conditions, while another feature would 'show patients everything from their nearest pharmacy to the best hospital for heart surgery across the country, with patients able to choose based on their preference'. These features will reportedly be introduced within the next three years, with an extra £10 billion allocated by Rachel Reeves in her spending review to fund NHS technology.
How exactly will NHS providers be 'inspired' by patients being offered choice?
Given the impressive capabilities of freely available AI tools, the need to spend £10 billion on bespoke NHS software appears questionable from the outset. Streeting's optimism about the NHS's skill with large IT projects is impressive. I have the NHS app on my phone, and it reliably opens without crashing. With a finger I can ask it to arrange repeat prescriptions or to show me my GP records, although in both cases it responds by saying it can't help as it can't connect to my surgery. The app also offers to transfer me seamlessly to 111 for medical advice where, after only five or six pages of warnings, it asks permission for it to pass on my medical history. With that granted, it passes me to 111 which, having received all my information, starts by asking me my sex at birth.
£10 billion, coincidentally, is the amount the Public Accounts Committee said, in 2013, had been wasted on an abandoned attempt to introduce an electronic patient record system to the NHS. 'The biggest IT failure ever seen', Dan Poulter, back then a Conservative health minister, condemned Labour's incompetence and responded with 'a £1 billion technology fund to help the NHS go paperless by 2018'.
Perhaps readers will be unsurprised to hear that the NHS is not, in fact, now paperless. Or to hear that we're already using artificial intelligence without the government's help. Some colleagues are shy about their use of ChatGPT for clinical knowledge but it's widespread and so it should be. Medical knowledge expanded beyond the bounds of a single person's memory generations ago, and knowing the best ways to look things up has been a key skill ever since.
For patients, ChatGPT, Gemini, Grok, and others already help. I used to sometimes advise people not to look certain matters up, confident that a quick Google would yield them the wrong end of a needlessly frightening stick. Today's large language models are better. For all their hazards, they respond to clinical questions not only with a decent chance of accuracy, but also with a high degree of useful context. They are the best products of some of the world's finest – and most highly paid – minds. Our NHS IT experts may well be poised to do a better job, but history suggests they are more likely to take ChatGPT and ineptly make it worse, wasting another £10 billion in the process.
Notably, Streeting's Blackpool speech was not chiefly about technology. He introduced the NHS App's new features as methods for tackling his real focus of inequality. Beveridge, in 1942, spoke of the need for a welfare state to fight the five giants of idleness, ignorance, disease, squalor and want. Streeting attributed today's inequality to 'poverty, a lack of good work, damp housing, dirty air, and the sporting, travel and cultural opportunities which are afforded to the privileged few being denied to the many'.
The shift worth noticing, because it is society's and not just Streeting's, is that people, and particularly the less fortunate, have ceased being spoken of as though they possess responsibility or control. Both, instead, rest wholly with the state. Streeting is right to care most for the least fortunate, whose opportunities are most constrained, but encouraging them to believe that what freedom they do possess is an illusion, and that they are not active agents but helpless victims, is harmful. Compassion, when sufficiently misguided, can be unkind.
When it comes to My Choice, the app's feature to 'democratise' the NHS, Streeting's thinking seems equally flawed. 'If NHS providers know that their waiting times, health outcomes of their patients and patient satisfaction ratings will all be publicly available,' he said in his speech. 'They will be inspired to respond to patient choice, raise their game and deliver services that patients value.'
This sounds attractive, just as it does to hold the state responsible for everything that's wrong in the world, but it makes as little sense. How exactly will NHS providers be 'inspired' by patients being offered choice? Here in the NHS almost everyone has a job for life, and equality means that within each field we are paid the same, regardless of talent or industry. And the senior managers, the only ones whose jobs are tenuous, almost invariably seem to fail upwards, leaving one post where they've been harmful in order to take another at a higher rate. Odd, to invoke the invisible hand of market forces in a context that effectively bans them.
Streeting appears to be a bright & decent man genuinely trying to improve the NHS. Against an admittedly dire collection of colleagues, he seems to be the highlight of Labour's benches. It is hard not to root for him, especially as any success he has will be our success too. Harder still, sadly, to see him pulling it off.