‘AI can’t just be overlaid on a failing health service’

Hetan Shah, chief executive of The British Academy, said that the NHS and UK government “can’t just overlay tech on a failing service”.

In a keynote speech at The Health Foundation event ‘Hype, hallucination, hope – what might AI mean for our health?’ in London on 13 February 2025, Shah said that policymakers “sometimes look at AI as if it’s magic”.

“I think that there’s a risk of a slightly naive view that technology will just save the day.

“But my question is, why would it be different this time when the NHS and public services are limited with failed digital transformation projects?” Shah said.

His talk follows the launch of the government’s AI Action Opportunities Plan in January 2025, which prime minister Sir Keir Starmer said would make the UK “an AI superpower”.

The plan includes creating a national data library, planned to unlock the value of public data and support AI development.

Referring to data as “the unsexy cousin of AI”, Shah said that AI’s quality  “is very much constrained by the quality of the data that it’s trained on and the UK has got better data than many other places, including the US”.

He added that primary care data is a “real problem” in terms of linking to the rest of the system.

“You need to solve the issue of GPs currently being data controllers,” he said.

Shah referenced Professor Cathie Sudlow’s independent review, published in November 2024, which called for UK health data to be treated as critical national infrastructure.

“I think if you look back during the pandemic, one of the things that was really striking was that the political will was there to make data much more accessible to enable it to be linked up and so on.

“So how can we get some of that energy back?” he asked.

Shah also called for robust evaluation of AI products to screen out the “snake oil” and “grifters” which he said were inevitable in a lucrative market.

“How do we know that somebody is not just relabelling a simple algorithm as AI, which has certainly been happening, or that the products are just not very good, as you’ve seen from multiple apps on medical and app stores?

“The trouble is it’s just very hard to get independently verified evidence in the space.

“You have a mixture of hyped press releases from the companies themselves, and, on the other hand, newspapers laying into the faults that you’ve seen in particular models, but it’s very hard to get a kind of actual, genuine evidence base,” he said.

Despite this Shah said that he is not pessimistic about AI and emphasised its potential to reshape the power dynamics in the health system and address determinants of health such as the environment, housing, education, transport and poverty.

“Thinking about how AI is going to play out across all of these domains is going to be really critical in terms of thinking about the impact on our collective health and wellbeing,” he said.