Is AI really revolutionising health care?

By day, Dr Philip Scott FBCS is Professor of Digital Health and Care at the University of Wales Trinity Saint David, where he runs a master's course in digital transformation in healthcare. Within BCS, he’s finishing his second term as chair of the BCS Faculty of Health and Care. His career journey, he explains, started with technology rather than clinical practice. He joined the NHS in 1994 and, over time, became more involved with clinicians and understanding what they needed from information systems.

The use of computers and information systems in healthcare, Philip explains, isn’t a new idea. Though the BCS Health and Care Faculty itself was only created a year or so ago, BCS has had a health and care group since the 1960s.

Describing the BCS Health and Care Faculty, he says: ‘Its purpose is to raise awareness of professional standards and increase participation in the professional body’s activities. That's things like education outside of formal school or university education… we help keep members up to date with what's happening in the industry — in the UK and globally. It’s about learning from other people. It’s also a human thing. It’s about making connections — friends.’

The impact of the on technology in healthcare

After five years in the role as chair of the Faculty, Philip is stepping down. And, it’s fair to say that a lot has changed in his industry in that half decade, particularly so because, when he took on the role, COVID-19 hadn’t begun.

‘The pandemic made things that were pilot projects into the norm and common practice overnight,’ he says, casting his mind back. ‘We used to talk about telemedicine — clinicians having video calls with their patients — and now we just call it medicine. And it has been recognised that it’s better for some patients; they don’t have to spend time waiting in hospitals and GP surgeries. The pandemic showed us that traditional things, the way we’d always done things, didn’t need to be written in stone.’

COVID’s acceleration of digital transformation has also been felt inside hospitals and by clinicians. Previously, gathering together a multidisciplinary team to collaborate on a complex cancer case would have involved much more diary synchronisation and organisation. Again, today, this can be done online, allowing teams to work more efficiently with increased agility.

‘This experience has made us appreciate the art of the possible, and in fairly simplistic things too. It changed people's awareness and their willingness to use technology. From virtual meetings, people have started to think about other processes’, he says, ‘and it’s made us think about re-use of data too. You capture data at the point of care and for that particular episode of care. Maybe that data can be looked at again to help understand service performance and even to make biomedical discoveries by re-using it at population level.’

How will AI impact healthcare?

Let’s be honest: when it comes to working with facts, popular generative AI tools have a well-earned reputation for being somewhat indiscriminate. This problem is compounded by these tools’ apparent self-assurance. Whatever subject you challenge your favourite chatbot to pontificate about, its response will be chipper, glossy and delivered with more confidence than a pedigree peacock in full display. Answers are always doubt-free, definite and definitive. 

For you

Be part of something bigger, join BCS, The Chartered Institute for IT.

Here’s a case in point. If you ask AI — in this case ChatGPT — how AI is changing digital health and social care, it won’t hold back: ‘Artificial Intelligence (AI) is no longer just a buzzword echoing in tech corridors — it’s a transformative force reshaping how we experience health and social care. From early disease detection to personalised treatment plans and streamlined caregiving, AI is rewriting the rulebook of care delivery.

Artificial intelligence is right here, right now and it’s helping us live longer, healthier, more independent and fuller lives. But is it? Or is this just another case of AI confidently hallucinating its way to an article’s end?

If you don’t like spoilers, look away now. According to Philip, AI isn’t ‘rewriting the rulebook’.

How AI can support healthcare

AI’s relationship with healthcare dates back to the 1980s. Back then the AI community’s focus was on capturing existing human knowledge and encoding it as an algorithm.

'The big problem with that approach was maintaining the algorithm as knowledge changed’, Philip explains. ‘[Algorithms] were very rule-based, and knowledge maintenance was the challenge.’

‘[What it is doing now is] making people think about what can be done differently. Most of the immediate benefits of AI in healthcare are process benefits — that’s admin as opposed to clinical work.’

For example, AI is helping move information from one service’s system to another department or organisation’s database. Traditionally, because services use their own systems that aren’t always interoperable, humans have hand-copied data from one to the other. Along with being slow, this opens up the risk for errors.

‘This process can be replaced by software, which is a form of AI’, Philip says. ‘It’s a crude interoperability… in computer science terms, it's clumsy. But, in process terms, it does things so rapidly, and that’s a big advancement. We’re bridging gaps between services.’

Looking forward five years, Philip believes this bridging use of AI may well continue and grow within the NHS. That’s because the service is full of legacy technology, and replacing these systems would be much more expensive than bridging them together with AI.

‘I can see AI being used in smart scheduling for waiting lists too’, he says. ‘And in managing operating theatres. So we don’t need a human constantly changing orders and priorities. Those are the sorts of places where an automated agent would be useful.’

AI in image analysis and diagnostic support

The main exception to the use of AI in admin over clinical usage, Philip explains, is in image analysis. Here, there has been a great deal of research and many advancements made in detecting tumours and making early-stage diagnoses.

‘This is routine now, but we still have humans in the loop’, he says. ‘If there is a positive identification, then a human steps in. It’s much the same as hallucinations in LLMs; there’s computation error in image identification.’

AI could also begin to make itself felt in the world of clinical decision support. Philip says: ‘There could be use cases where the pathway to a diagnosis is half an algorithm and half clinical judgement. Something that could present a logical workflow to a diagnosis would be a big advantage. [A patient’s history] might lead to a differential diagnosis... so, possibly, the AI could suggest a test or even automate requesting that test. AI could become an iterative part of the diagnostic process.’

Drawing the conversation to a close, Philip says: ‘The architect Clough Williams-Ellis, famous for Portmeirion Village, said: “Cherish the past, adorn the present, construct for the future." That sums up my attitude [to technology]. Learn from the past, make the present better and build for the future. I love that.’