Nicolas Brault, associate professor in human & social sciences on the Beauvais campus, has just published his latest scientific article dealing with the critical evaluation of artificial intelligence in health.
Available in the Journal of Evaluation in Clinical Practice, this paper focuses on mobile applications dedicated to this universe, and the processing of data collected by these applications.
Author's abstract :
Artificial intelligence and big data constitute a revolution in the medical field.
Coupled with mobile health applications (mHealth) that everyone can use on their cell phone or smart watch, these technologies are paving the way for so-called "personalized" medicine, which can be adapted to the individual characteristics of each patient, in terms of prevention, diagnosis and treatment monitoring, particularly for patients suffering from chronic diseases.
However, we consider that personalized medicine is more of a technological myth than a clinical reality. Indeed, this article intends to demonstrate that what patients and physicians need is not more data but a higher quality of data, in particular to reduce or avoid bias. Indeed, it is imperative to critically evaluate not only the validity of these data, but also the inferences made by algorithms based on these data. Only humans are capable of critically evaluating algorithmic tools for data analysis.
Against the thesis that artificial intelligence would replace physicians in the future, we therefore believe that the human dimension must remain at the heart of the doctor-patient relationship, both from an epistemological and an ethical point of view.
