의사의 설명의무와 인공지능: 글로벌 시사점 및 법정책 방향을 중심으로

Modern medicine is undergoing revolutionary changes in diagnosis, treatment, and prognosis prediction due to the rapid advancement of artificial intelligence (AI) technology. As AI assists or partially replaces physicians' judgments, new legal and ethical challenges are emerging concerning the physician's duty to explain—a core means of realizing patients' self-determination rights. This paper empirically analyzes the current status of AI adoption in healthcare and examines the legal nature of the duty to explain under medical law, focusing on constitutional and contractual grounds. It reviews the approval status of medical AI devices domestically and internationally, the adoption of AI in healthcare such as generative AI for automated medical record generation, and conducts a comparative analysis of overseas legislative examples and international organization guidelines. The research findings indicate that AI replacing the duty to explain faces structural limitations. The duty to explain is fundamentally about patient-centered communication and empathetic interaction. However, current AI struggles to embody these human elements, and the technological reliability of deep learning-based AI is insufficient due to its black-box nature and algorithmic bias. Furthermore, replacing the physician's duty to explain with AI could lead to legal and ethical issues such as infringement on the patient's right to self-determination, a gap in responsibility, and a decline in the physician's judgment capabilities. Therefore, it was concluded that AI cannot fully replace the duty to explain at this stage and should only be used as an auxiliary tool under strict conditions. Consequently, it is necessary to codify the principle of AI as an auxiliary tool through amendments to the Medical Service Act, mandate disclosure to patients about AI usage and potential errors, and strengthen explainability standards through regulatory enhancements. Additionally, a phased introduction strategy based on risk levels, expanded AI training for medical professionals, and the establishment of a responsibility allocation system between physicians and manufacturers are required. Legal policies for the era of medical AI must pursue a balance between technology and humanity, innovation and safety, by codifying the core principles of ‘final supervision and responsibility by human physicians’, ‘guaranteeing patients’ autonomous choice', and ‘transparent disclosure of AI limitations’. Centering on the ultimate goal of healthcare—patient health and well-being—we anticipate that artificial intelligence will be utilized as a beneficial tool for both patients and medical professionals.