Index of “Ethicality” of AI Systems in Medicine: From Theory to Practice

Abstract

The article presents the methodology developed in the HSE University – Index of Ethics of Artificial Intelligence Systems. The task of developing this Index was to assess real and possible ethical risks arising at all stages of the life cycle of AI systems. The system itself does not possess any “ethics”, while socially acceptable, morally permissible, and necessary may be the actions of developers and data providers in the process of its design, as well as of operators and consumers in the process of piloting and implementation. Several issues re­lated, for example, to the confidentiality of personal data, to liability, are partly regulated by the current legislation, while most of the threats are only predicted. The methodology is de­signed for the purposes of the medical community, which needs to confirm the moral sound­ness and compliance with professional standards for the introduction of new technologies into clinical practice; and as a tool that can be used in the activities of ethics commissions, similar to bioethics committees, when approving relevant medical research involving AI; as a supplement to voluntary technical certification procedures, as well as in forensic medi­cal examinations. The Index reflects the most pressing issues – such as issues of trust, dis­tributed responsibility, data privacy, transparency and explainability of AI models, fairness, and non-discrimination. The uniqueness of the Index lies in its underlying interdisciplinary approach, which includes the methodology of “value design” and field sociological research, which made it possible to combine theoretical humanitarian approaches to understanding the content of key moral concepts, with their interpretation in professional medical ethics. This makes the “Index” not only interesting from a theoretical point of view to formalize ethical concepts, but also useful from a practical point of view – as an example of the ap­plied meaning of ethics and the use of its tools to solve socially significant problems.