Artificial intelligence in medicine: some – significant, but partial – answers offered by the code of medical ethics (on non-discrimination, informed consent and the doctor-patient relationship)

Abstract

This study aims at assessing the impact produced by the use of artificial intelligence systems in medicine in relation to some aspects of primary importance in medical practice. In particular, the object is to highlight how the use of medical AI systems: increases the risk of discriminating in the provision of healthcare services, as a result of the use of biased data and incomplete data in the training of machine learning systems; undermines the informed consent, in view of the need to provide complete information also concerning the decision-making logic of the computer systems that may be involved in the diagnostic or therapeutic activity (all this bearing in mind the well-known forms of opacity of AI systems); progressively transforms the doctor-patient relationship with effects on the nature of clinical evaluations and on the role of the doctor, who is called upon to act increasingly as an intermediary between the patient and the computer system. Starting from these premises, the study will attempt to assess the adequacy of the solutions to the mentioned issues that can be offered by the Code of Medical Ethics currently in force.