ARTIFICIAL INTELLIGENCE IN MEDICAL PRACTICE: REGULATIVE ISSUES AND PERSPECTIVES

Abstract

The aim of the research is to identify specific of AI in healthcare, its nature, and specifics and to establish complexities of AI implementation in healthcare and to propose ways to eliminate them. Materials and methods: This study was conducted during June-October of 2020. Through a broad literature review, analysis of EU, USA regulation acts, scientific researches and opinions of progressive-minded people in this sphere this paper provide a guide to understanding the essence of AI in healthcare and specifics of its regulation. It is based on dialectical, comparative, analytic, synthetic and comprehensive methods. Results: One of the first broad definitions of AI sounded like “Artificial Intelligence is the study of ideas which enable computers to do the things that make people seem intelligent ... The central goals of Artificial Intelligence are to make computers more useful and to understand the principles which make intelligence possible.” There are two approaches to name this technology - “Artificial intelligence” and “Augmented Intelligence.” We prefer to use a more common category of “Artificial intelligence” rather than “Augmented Intelligence” because the last one, from our point of view, leaves much space for “human supervision” meaning, and that will limit the sense of AI while it will undoubtedly develop in future. AI in current practice is interpreted in three forms, they are: AI as a simple electronic tool without any level of autonomy (like electronic assistant, “calculator”), AI as an entity ith some level of autonomy, but under human control, and AI as an entity with broad autonomy, substituting human's activity wholly or partly, and we have to admit that the first one cannot be considered as AI at all in current conditions of science development. Description of AI often tends to operate with big technological products like DeepMind (by Google), Watson Health (by IBM), Healthcare's Edison (by General Electric), but in fact, a lot of smaller technologies also use AI in the healthcare field – smartphone applications, wearable health devices and other examples of the Internet of Things. At the current stage of development AI in medical practice is existing in three technical forms: software, hardware, and mixed forms using three main scientific-statistical approaches – flowchart method, database method, and decision-making method. All of them are useable, but they are differently suiting for AI implementation. The main issues of AI implementation in healthcare are connected with the nature of technology in itself, complexities of legal support in terms of safety and efficiency, privacy, ethical and liability concerns. Conclusion: The conducted analysis makes it possible to admit a number of pros and cons in the field of AI using in healthcare. Undoubtedly this is a promising area with a lot of gaps and grey zones to fill in. Furthermore, the main challenge is not on technology itself, which is rapidly growing, evolving, and uncovering new areas of its use, but rather on the legal framework that is clearly lacking appropriate regulations and some political, ethical, and financial transformations. Thus, the core questions regarding is this technology by its nature is suitable for healthcare at all? Is the current legislative framework looking appropriate to regulate AI in terms of safety, efficiency, premarket, and postmarked monitoring? How the model of liability with connection to AI technology using in healthcare should be constructed? How to ensure privacy without the restriction of AI technology use? Should intellectual privacy rights prevail over public health concerns? Many questions to address in order to move in line with technology development and to get the benefits of its practical implementation.