We are interested here in the possibilities offered by Artificial intelligence (AI) technologies in the health care sector and in particular in supporting medical decisions. Combining AI technologies with clinical decision support systems enable to provide clinicians with new information that was impossible to generate instantly till now. Indeed, it make it possible to estimate the impact of patients' clinical and socio-demographic characteristics in terms of effectiveness (benefit/risk balance) and efficiency (incremental cost/effectiveness ratio (ICER)). For example, patients' age, their health histories and comorbidities may have an impact on both the expected benefits of treatments (e.g., life expectancy and quality of life gains, risks of complications or adverse effects, etc.) and on costs (e.g., hospitalisation costs in case of serious health events, costs of adjuvant treatments, etc.). So far, economists have estimated average ICERs for target populations, also called "populations of indication" because the only available efficacy data are average efficiency data. First, we address the question whether estimating individualized ICERs would be consistent with the classical utilitarian framework, as well as with Paretian welfare economics framework. Theoretically, estimating individualized ICERs would make it possible to better attain the objective of maximizing utility associated with health care under budgetary constraints. Second, we address the question of the social acceptability of medical decisions based on an individualized ICERs. This would imply that a treatment for a given indication could be recommended for some individuals, but not for others. AI based-clinical decision support systems might reinforce implementation of utilitarian justice models by overcoming the current limitations pertaining to the information capacities of the actors. These avenues may therefore raise ethical controversies, that are maybe not unprecedented but certainly intensified. Thereby they make it even more necessary to organise institutional discussions on the ethical frameworks that must be favoured.