Will artificial intelligence revolutionize the medical profession?

Coined in the 1950s, the term artificial intelligence (AI) refers to a computer system capable of reasoning, learning and planning, with behavior reminiscent of intelligent biological systems. Its automatic learning, or machine learning, relies on programming that is not only determined by a code, but also adapts to the data it is provided with.

Matthias Brunn, University of Montpellier and William Genieys, Sciences Po - USPC

For example, within the formal framework of artificial neural networks, mathematical algorithms enable us to iteratively find the best way to represent the link between several variables. Each neuron represents a fairly simple function, making it possible to construct highly complex functions in a large network.

Medicine, for its part, is an ancient and recognized institution in our societies, as demonstrated by the trust placed in healthcare professionals during the Covid-19 pandemic. And we tend to regard it as a great ship whose stability cannot be threatened.

But the reality is quite different. In the decades to come, medicine is likely to be profoundly transformed by AI, a veritable "revolution" in the words of President Emmanuel Macron. And to get the measure of these coming upheavals, we need only refer to the Health Data Hub, a platform launched by the French government in November 2019, enabling researchers to use data from the national healthcare system to train artificial intelligence models.

Some 10 research projects have already been selected to take advantage of this platform. And all of them aim to improve the effectiveness of medicine. For example, "Deepsarc" aims to identify the best therapeutic regimens for the treatment of sarcoma, "PIMPON" aims to help prescribers better understand dangerous drug interactions, "Hydro" seeks to predict heart failure attacks for pacemaker patients, "Oscour" aims to mobilize emergency room data to improve health surveillance, "Deep.Piste" will evaluate the contribution of artificial intelligence to organized breast cancer screening - a major contribution, according to a study of six radiologists published last January in Nature.

While the medical profession has always been able to take advantage of technological innovations (X-rays, MRIs, etc.), it is now faced with an unprecedented situation of disruption. Will algorithms overtake healthcare professionals? Within which disciplines will battles be fought?

A world in constant evolution

According to American sociologists Rue Bûcher and Anselm Strauss, any profession can be defined as a world made up of segments, with different conceptions of their practice, and varied identities that transform, maintain, develop and/or disappear. Various medical disciplines thus find themselves in competition. For example, when it comes to the management of serious coronary disease, there may be opposition between cardiologists who advocate angioplasty and stenting, and cardiac surgeons who prefer open-heart bypass surgery.

In the ^twentieth century, with the support of the state and in competition with other groups such as spiritual healers, the medical profession as a whole gained a monopoly on the provision of health care. To achieve this, it relied on a coherent, scientific system for defining, classifying and diagnosing phenomena of somatic and/or psychic malaise). But it is also its ability to self-organize (into disciplines, learned societies, unions, etc.) that has given medicine its authority in health matters.

In such a context, AI can be seen as a new technical innovation capable of helping doctors, but also of reinforcing their power vis-à-vis other disciplines. Yet within this highly hierarchical profession, technological choices are made by elites who operate within a "mandarin" logic, based on a mix of conservative professional values and rejection of the bureaucratic power of the state. Even if, for each discipline, the decision to use AI depends on a comparative assessment of its benefits and risks, based on past experience.

Psychiatrists, for example, are rather skeptical that AI will be able to perform the complex tasks they perform on a daily basis, as psychiatry is generally less fond of technology than other disciplines. Conversely, medical imaging is a field where the performance of AI is already making certain radiologist profiles obsolete. In fact, we are witnessing the emergence of a first generation of radiology professors who have built their careers on the use of AI in medical imaging.

Strong disruptive potential

By moving beyond traditional technical innovations, AI is confronting medical professionals with a veritable revolution. The reasons are primarily technical: in addition to unprecedented speed in the execution of certain tasks, it enables the development of autonomous learning - as demonstrated by the lightning progress of instant language translation, now offered free of charge on the web, with major implications for the job market. But the revolution also comes from the anthropomorphic side of AI...

Artificial neural networks, robots with a human face and chatbots are all examples of this. These digital conversational agents use AI methods via text and/or voice to mimic human behavior with evolving dialogue. And they are seen as a means of providing mental health care in areas where little exists, or for people who have difficulty revealing their feelings to a human being. In short, AI itself becomes a social actor (or an "actant" in Bruno Latour's universe), increasing its disruptive potential.

These two facets of AI go hand in hand, and risk profoundly transforming the world of medicine by upsetting the established order. For example, radiology could become a specialty with fewer positions, and a profile that is more "clinical" than "diagnostic". On the other hand, psychiatry, which currently attracts few students, is likely to see its profile strengthened due to its social dimension, which is difficult to substitute with AI applications.

In fact, it is precisely by building on this dimension and its founding virtues (empathy, intimacy of the patient-practitioner relationship, medical confidentiality) that the medical profession as a whole could redefine its profile and scope of action. Indeed, in the field of healthcare, a number of players outside the medical world have positioned themselves in the care of somatic and/or psychological malaise through AI.

GAFAM like Google, which funds studies on AI in healthcare, Facebook, which mobilizes AI to detect the risk of suicide among its users, or Microsoft, which in France hosts the Health Data Hub.

Like a container ship that, although accustomed to storms, can be battered by a surge of rogue waves, the medical profession is facing an unprecedented situation. And even if it has the necessary resources to navigate, the power of the elements may force it to find a new course. Whatever the outcome, the transformation will be profound.

Matthias Brunn, Researcher in political science at CEPEL/CNRS, University of Montpellier and William Genieys, CNRS Research Director at the CEE at Sciences Po, Sciences Po - USPC

This article is republished from The Conversation under a Creative Commons license. Read theoriginal article.