The USE of artificial intelligence technologies in the electronic healthcare system of Ukraine

Abstract

The article examines the legal and organizational aspects of the use of artificial intelligence technologies in Ukraine’s electronic healthcare system and the electronic medical information systems integrated with it. The principal areas of application of these technologies are outlined, including the processing of electronic health records, medical image analysis, disease risk prediction, personalized treatment, support for telemedicine services, healthcare resource management, and cyberthreat detection. The study analyses the risks associated with insufficient data quality and representativeness, algorithmic bias, the lack of transparency of automated outputs, breaches of medical data confidentiality, unauthorized access to personal data, and uncertainty regarding legal liability. It is emphasized that artificial intelligence systems should perform an auxiliary function and must not replace the professional clinical judgment of healthcare practitioners.

The methodological basis of the study comprised formal-legal, systemic-structural and comparative-legal methods, as well as methods of analysis, synthesis and generalization. Their application made it possible to analyse national and European legal regulation, determine the place of AI systems in the electronic healthcare infrastructure, and substantiate proposals for a risk-based model of their integration.

A risk-based model of legal regulation is proposed, under which the scope of applicable requirements should depend on the intended purpose of the system, its degree of autonomy, the nature of the data processed, and the potential consequences for patients’ lives, health, and rights. Such a model should provide for the classification of systems according to their level of risk, technical and clinical assessment, auditing of the quality and representativeness of training data, transparency and explainability of algorithmic outputs, meaningful human oversight, logging of operations, and continuous monitoring.

A phased procedure for integrating artificial intelligence technologies into the electronic healthcare infrastructure is formulated, and the establishment of a register of systems authorized for use is proposed. Approaches to the allocation of liability among healthcare practitioners, healthcare institutions, developers, software providers, and entities responsible for operating electronic medical information systems are identified. The principles of legality, human-centeredness, data minimization, non-discrimination, safety, transparency, accountability, and traceability of automated operations are specified.