AI in the NHS: A Promising Future, but No Silver Bullet

Over the course of recent weeks and months, parties of all political shades have spoken about the need to “embrace” and “harness the potential” of artificial intelligence (AI) within the NHS. As political parties prepare to release their general election manifestos, I expect there to be more of the same. Ahead of those manifestos being released, I want to issue a word of caution.  

AI has great and exciting potential but will not solve all the current challenges faced by clinical radiology and clinical oncology, or indeed the wider NHS. 

NHS waiting lists will not magically disappear because of AI implementation. It needs to be carefully embedded as part of a wider set of measures including supporting the workforce which is already overstretched. Poor implementation of digital solutions can even mean that doctors have less time to spend with patients, so it is vital that we get this right.

The RCR’s specialities, clinical radiology and clinical oncology, are technology-centric and thus are at the forefront of the AI revolution. Many of you will have seen the news coverage of AI in the detection of cancer from breast mammograms, an exciting development which has the potential to save lives through early detection. We are therefore cognisant of the opportunity to leverage these new AI tools to deliver maximum benefit.  

This will not, however, just happen automatically. AI deployment is a complex beast, with many factors to consider when undertaking such projects. Some AI tools are already business-as-usual, humming away in the background and hardly noticed most of the time. But other tools – the game-changing items – require more work to get up and running. 

Given the pressures that colleagues across radiology and oncology are under, with decreasing levels of retention and logjams in the training pathway leading to increasing workforce shortfalls, it is vital that we embrace any innovations that could boost capacity. AI presents the opportunity for capacity to be boosted, but only if the NHS can seize it.

As this significant interest from policymakers and politicians intensifies, it is worth remembering two things.  

Firstly, that AI will not and should not replace highly trained medical professionals, but rather can empower them to further help their patients and focus their care. Many processes and pathways are not fully automatable, and arguably nor should they be. Patients want to know that there is a human doctor overseeing their care. What AI will do, though, is change the ways in which we work – so we need to be ready. 

Secondly, we need to get the basics right first, if AI is to be in any way the revolutionary change which many have professed it will be. It is no good having a sophisticated AI tool at your disposal if your patients miss their appointments because they receive invitations via letters which arrive through their door after the date of their appointment. It will not help your working lives to have advanced auto-contouring or image analysis tools if it still takes your computer 20 minutes to turn on in the morning. And it is no good expecting the NHS to undertake an enormous programme of change if it does not have enough staff – both clinical and non-clinical – to make that change happen. 

The RCR speaks regularly with NHS England, the Department of Health and with other relevant organisations to share our recommendations and to shape AI policy. This has included meetings with senior government officials in England as well as the devolved nations. Recently, we submitted evidence for the Science, Innovation and Technology Committee Inquiry on Artificial Intelligence governance. We will continue to put forward the views of our Fellows and members in these fora as we seek to work collaboratively with government and the NHS on this increasingly important topic.

So, in the week that election manifestos are published, I am calling on all political parties, whoever forms the next government, to adopt a stronger and more coordinated effort from across government and state bodies to deploy AI into the NHS comprehensively, safely and effectively.  

I look forward to working collaboratively and constructively with the next government to make this a reality. I implore all parties to remember, when calling for the NHS to embrace AI and new technologies, that it will not make the significant challenges our workforce faces on a daily basis disappear overnight.