[Artificial intelligence systems and precision medicine: hopes and realities.]

Precision medicine (MP), using machine learning (ML) techniques of artificial intelligence (AI), analyzes the so-called "big data" to improve diagnostic skills and predictive response to therapy, in order to tailor the treatment on individual characteristics. Furthermore, the use of large databases is limited by their very nature, as data and not values, capable of threatening the validity and generalizability of the conclusions. Currently, for this reason, interventions on populations, subpopulations or groups are still fundamental, favoring, especially in the pharmacological field, the response of the "average patient", over the particular case. Therefore, it is essential to carry out methodologically robust, prospective studies of comparison between teams of doctors who use systems based on algorithms and teams that do not are using them, avoiding direct comparisons, less significant, between doctors and AI. Only in this case will technologies effectively applicable in real rather than experimental clinical contexts be implemented in the near future.