AI techniques -- and more specifically machine learning -- applied to the field of health give rise to hopes of improvements in patient care: more accurate diagnoses, better choice of treatments, better organisation of cares, etc. At the same time, there are concerns about their nature and their use. We are particularly interested in the use of AI as a tool to make diagnosis, prognostic or prescription in healthcare. In this, we are particularly interested in AI in the context of epidemiology. An important question is to identify in what way AI, as a technical object, raises or not new legal or ethical issues. To do so, we will be interested in highlighting the different issues that arise during the life cycle of AI from its conception to its use; and in highlighting different approaches -- preventive, prescriptive, corrective -- that regulate these issues. Finally, we question the possibility of integrating legal or ethical regulations into the design of AI.