AI can assist in the recovery of humane medicine, writes Dr Stephen Davies.

The last year has seen a surge of often excited discussion around AI and its potential.

One of the key breakthroughs is the production of so-called generative AI: AI that does not only accumulate and order information either as an end in itself or to deliver a specific and defined task, such as playing chess, but also creates new things using that data. This often involves extensive interaction with humans to prompt the productive process. The interaction is itself an important driver of the process of learning, by which AI increases its capacity and refines and improves the quality of its output.

Most of the attention has focused on the impact of generative AI in areas such as art, music and the production of written texts, but it has far wider applications. One of the most important of these is medicine and healthcare.

Generative AI can be an empowering resource for individual patients. The AI algorithms will be able to access the entirety of what is known about healthcare, illness, human biology – and the patterns of these among human populations – and will be able to do so in an instant.

The emergent models of AI will also be able to conduct an examination or assessment of the patient, partly through responses to questions and partly through tests and measurements. This means that, without going to a doctor or other practitioner, a patient will be able to generate an entire health profile for themselves. This would include identifying any possible future health complications.

It also means that AI will be able to diagnose presenting problems or issues and, because of its greater and swifter access to data, it will be able to do so with greater accuracy than most human physicians.

What all of this means is the empowerment of the patient. Each individual patient could have access to what is effectively a personal physician and health advisor.

All of this suggests that AI will bring about a transformation in the organisation and delivery of healthcare. As well as giving greater autonomy and independence to patients, it can also assist the practitioners. At first sight, this means far fewer doctors and other practitioners – much of the work currently done by GPs can be done by AI.

AI will also accelerate the move away from large integrated hospitals to a model of a much larger number of smaller clinics where many routine procedures will be performed, often by AI rather than by humans. Hospitals are likely to revert to their two classic functions of emergency care and convalescence – the latter being something that has declined massively in the last few decades due to pressures on hospital bed numbers in the UK.

All this might suggest a future in which the human element in medicine is removed in the name of efficiency, but in fact the opposite could be true. What AI has the potential to do is to reverse a trend that has been going on for decades and restore a humane medicine in which human relations are the central feature – but only if we make the right choices about how to use it.

Over the last few decades, medicine and healthcare – like many areas of life – has come to be dominated by elaborate rules and processes, to the diminution of human relations and choices. This is often blamed on technology, but it reflects the choice made to fit technologies into a particular kind of structured managerial system rather than letting technology assist individual patients and practitioners in having greater independence.

AI should be thought of as complementing human choice and interaction rather than replacing it. What AI can do is to carry out most of the routine and mechanical aspects of healthcare more efficiently while at the same time distributing decision making more widely by empowering both patients and lower grade practitioners.

With all that said, there are two things that AI cannot do. First, tasks that require dexterity; because of the high cost and technical challenges of replicating the combination of human eye, hand and brain. This includes most surgical procedures. Second, actions that involve human relationships. Even if an AI passes the Turing Test, it is extremely unlikely to be able to replace this, because of the importance of physical connection. This means human doctors will be able to refocus on those two kinds of tasks.

In sum, when utilised effectively, AI can support practitioners and patients in the creation of healthcare as a social product. Right now, medicine is suppressed by a system of formal procedures driven by the bureaucratic parts of healthcare practice. This policy is wrong for all kinds of reasons, not least because the reality of human relations and the tacit knowledge of both patients and practitioners cannot be captured by formal rules.

AI will expand the autonomy of patients and doctors alike, freeing individual practitioners from the managerial system they are currently constrained and slowed down by. It could even take over much of the mechanical processes and free up the actual practice of medicine as a social relationship, while also enabling the localisation of what has become an increasingly centralised and large scale system, reimplementing humane medicine and a genuine focus on personalised care.

Dr Stephen Davies is the Senior Education Fellow at the Institute of Economic Affairs.

This article was published in the latest edition of Centre Write. Views expressed in this article are those of the author, and not necessarily those of Bright Blue. 

Read more from our June 2024 Centre Write magazine, ‘Generation AI?’ here.