To date, there is no established practice or unified approach to the issues of liability for errors made on the basis of machine learning algorithms. Typologically, the subject may be a physician or other medical workers, the head of a structural unit or medical institution, the developer of an artificial intelligence (AI) system, a robot or their components, as well as an information system operator and employees of a technical department. From the standpoint of risk taxonomy, the following can be distinguished: harm to life and health caused as a result of a diagnostic or prognostic error, a defect in the implementation or updating of a model, insufficient training or lack of qualifications when working with AI systems, compromise of an information system, violation of legally protected secrets, discrimination and unequal access to medical care, and others. This paper provides an overview of select AI solutions in the healthcare sector, relevant legal regulation, judicial and law enforcement practice in Russia and foreign countries, as well as the possible legal qualification of certain criminogenic situations.