Streeting’s Strategic Mistake
CARL HENEGHAN AND TOM JEFFERSON
There are a few clear reasons why UK Health Secretary Wes Streeting has been publicly and repeatedly talking about, promoting and prioritising Artificial Intelligence (AI) in the health service (the NHS).
Streeting has framed the NHS as under enormous pressure from chronic waiting lists, staff shortages and outdated systems, saying the service is “broken” and needs reform alongside investment. Introducing AI and digital technology is a key part of his efforts to transform the NHS into a more digital, modern, and efficient system.
Repeatedly, he has described AI as a tool that can help free up clinicians’ time by reducing bureaucracy and paperwork so that doctors can spend more time with patients-something he argues is crucial if the NHS is to cope with rising demand.
The Streeting approach, which currently lacks specific details and solid evidence, can be seen as politically motivated. He has incorporated AI into the government’s “Plan for Change” and the broader ten-year NHS strategy. His aim is for the public to view him as a leader who is proactively reforming health care rather than merely pouring money into existing issues.
He presents his strategy as part of a Labour government that is progressive, innovative, and focused on modernisation—an appealing pitch for someone aspiring to be Prime Minister.
Additionally, he wants the electorate to believe he is actively addressing NHS challenges rather than being passive, which is crucial given the public’s growing frustration with waiting lists and GP access.
To bolster his ambitions, an extra £10 billion has been allocated in the Chancellor’s spending review for NHS technology and digital transformation by 2028-29, aimed at enhancing the NHS App, single patient records, and digital access. However, some question whether this funding is new or merely reallocated.
Yet, expensive NHS tech projects have a poor track record, and AI plans risk repeating old mistakes rather than solving core problems.
A major NHS electronic patient record programme (NPfIT) was abandoned in 2011 after it cost taxpayers over £10 billion. The NPfIT promised exactly what today’s AI rhetoric promises: Efficiency, joined-up records, time saved for clinicians and better patient care. Instead, it delivered systems clinicians didn’t want or trust, contracts locked into unsuitable software and years of disruption with little benefit.
When Streeting talks about AI as “transformational”, those who have been around long enough immediately hear echoes of the same over-promising language used decades ago.
As a result of past failings, AI tools in the NHS should face drug-level scrutiny, not pilot-level enthusiasm. Claims about productivity gains must be independently validated, and pilots should be slow, reversible and tightly evaluated.
There you go again, asking for evidence!
What Streeting would need to do instead is start with small, narrow use cases (e.g., admin automation, specific diagnostic workflows); require clear published evidence of benefit before scaling; and accept that some pilots should fail and be stopped without political embarrassment.
Can ministers publicly abandon an AI project that doesn’t work? Probably not.
If not, the TTE office will assume history is repeating itself. Without mandatory independent clinical trials of AI tools, clear regulatory pathways, and continuous post-deployment monitoring, AI will remain hype rather than an adequate healthcare solution.
The NPfIT mistake was the implicit promise that tech could transform care on its own. History has a way of repeating itself, and Streeting is falling into the same trap. AI will not solve staff shortages, the social care crisis, or underfunding.
Streeting’s most significant strategic mistake so far is proposing AI as the “solution” to such workforce shortages. It isn’t that AI can’t improve productivity; it’s that saying it too loudly triggers defensive mechanisms. Clinicians hear replacement, unions hear cost-cutting, and the public hears automation over care. Even if unintended, this framing destroys trust, hardens resistance and invites comparisons to the £10 billion NPfIT program.
Currently, Streeting meets rhetorical tests better than operational ones: AI is being promoted faster than the evidence base justifies.
Successful NHS AI should start locally, be optional, save time immediately, and leave accountability unchanged. Most importantly, it should produce evidence before publicity – referring to what works as opposed to what might work.
Hyping AI as a “game-changer” risks repeating previous mistakes: grand national promises, pilots launched before proof, and technology sold as a substitute for people. Add the push toward apps and digital front doors, and it starts to look like algorithmic rationing, with the digitally confident sailing through while everyone else is left behind.
Older people, lower-income people, and people who don’t use their phones will be pushed to the back. The NHS was supposed to be universal, not best experienced with Wi-Fi.
AI can help if it stays quiet, local and firmly under clinicians’ control, but oversold, centrally driven tech projects are how the NHS ends up wasting money it can’t afford to lose, yet again.
Two old geezers wrote this post who believe that if technology helps healthcare quietly, keep it; however, if it needs a press release, be suspicious.
This article (The Government’s Obsession with Artificial Intelligence and Tech) was created and published by Trust the Evidence and is republished here under “Fair Use”
Leave a Reply