The Redesigning Health theme aims to explore through roundtables, discussions and research, trends in healthcare that focus on how health systems can be redesigned and centred on prevention and improving population health.  In other words, how can the health system in the UK focus on health and not just sickness; or at least attempt an equality between the long term, ie prevention and the short term, ie treatment. Concepts very much in vogue with Prime Minister Andy Burnham having talked about ‘building a more preventative state' and in the Budget the chancellor, John Healey, stating the ‘Budget…reaffirms the Government's commitment to reform through devolution, prevention and the transformational use of technology.'

This first roundtable, which Lee Peart reports back on below, looked at how AI and data have the potential to be the new engineering that can help prevent disease - in much the same way as Sir Joseph Bazalgette's sewerage system did for infection in 19th century London. AI is not new - machine learning has been used in health systems for many years. However, the recent rapid increase in its' interest and use brings with it both positives and negatives, and as with the broader use of health data for research, planning and innovation there are ethical implications which if not managed and planned effectively, could result in losing public trust. This tension between ensuring the health system can benefit from AI innovations, while not losing public trust, was a key point of discussion in the roundtable.

The last few weeks have seen a flurry of international activity concerning the governance and societal implications of the use of AI - much of which has centred on questions of trust and human agency.

In his speech at September's UN General Assembly, Prime Minister Andy Burnham called for a Global agreement in AI governance. This was followed later that month when Pope Leo IV  addressed UNESCO on his visit to Paris where he warned against the dangers of AI and said that new technologies must ‘remain at the service of the human person' rather than becoming ‘yet another instrument of domination and injustice'. Industry has also entered the debate, with Dario Amodei, chief executive of AI firm Anthropic, in an essay published in September, calling for a slowdown in AI capability gains to allow safety, alignment and governance to catch up. Also in September, in a Guardian interview, the Nobel Prize winner Geoffrey Hinton, credited as the ‘Godfather of AI', doubled down on recent warnings that AI models pose an existential threat to humanity and rejected suggestions that industry could regulate itself. Finally, the recently published Lancet Commission on 21st Century Global Threats to Health warned the ‘malicious use of AI, including bioweapons, and nuclear conflict with nuclear famine carry risks of extinction in the next 75 years'.

So what are the public's attitudes?  A report published last year by the Royal College of Radiologists found while just 40% overall say they feel comfortable with AI in healthcare, that figure rises to 76% among those who are very familiar with AI. However, results from an Office for National Statistics survey conducted in August 2025 found 32% of adults in the UK would trust healthcare services to use AI for some tasks. Notably this was down from 36% in November 2024. What is clear is the public point of view is fluctuating and probably won't take much to move opinion in either a positive or negative direction.

Other findings from the Royal College of Radiologists' survey build on the trust question, including the fact the public want human doctors to continue being involved, with AI being used to assist rather than replace doctors and that doctors should oversee AI and retain responsibility for patient care. In what, sadly, remains a continuing point of anxiety, the public want assurance their personal health information remains safe. Given recent stories of NHS staff breaching medical records of people involved in high profile incidents, it's not hard to see why the public remain concerned.

This point about AI being a tool to assist humans rather than replace them could conflict with what is frequently seen as a benefit of AI, that of productivity. In last week's Budget, the chancellor announced he is setting up an independent review into NHS productivity and AI that will inform the next spending review.

In a study published in May, University of Pittsburgh researchers concluded productivity didn't rise after companies invested in AI and cut staff. Although executives have been overwhelmingly optimistic about AI's potential impacts for years, ‘management optimism does not effectively translate into … better performance', the paper states.

There is no doubt AI has the potential for improving NHS productivity along with redesigning patient services, but as with the rapid speed in AI capability, rushing to the  - albeit perceived but not yet evidenced – ‘goldmine' that is AI productivity - could undermine trust. Trust needs to be earned and neither demanded nor expected; accountability for decisions, errors and mistakes should not be abnegated to the tech but remain with humans; AI should not be used to distance doctors from patients and decision makers from responsibility. In short, there is a stronger ethical objective in ensuring the public can trust the NHS to use AI competently, responsibly and with accountability than attempting to make the public more trusting of AI. To quote the bioethicist Professor Onora O'Neill: ‘Trust is the response. Trustworthiness is what we have to judge.'

The next roundtable in the Redesigning Health series in December, will take forward the discussion from this discussion and look at how innovations can be prioritised to improve prevention and the public's health - and at a time when resources constrained, how a multi-sector approach can support investment.

Christopher Exeter is a global researcher in healthcare.