「인공지능기본법」시대 의료 인공지능의 활용을 위한 제언

This Article examines the regulatory structure and legal implications of medical artificial intelligence under the Framework Act on Artificial Intelligence, enacted in the Republic of Korea in 2025, with particular emphasis on medical AI systems and AI-enabled medical devices. Historically, medical AI in Korea had been regulated only indirectly through the Medical Devices Act, personal data protection statutes, and bioethics-related regulations. However, the enactment of the AI Framework Act introduces a general, cross-sectoral legal regime governing artificial intelligence as a whole, thereby signaling a structural transformation in the legal status and regulatory approach to medical AI. Against this backdrop, the Article analyzes the legislative purpose and normative structure of what is known as the “Artificial Intelligence Frame- work Act” and evaluates how medical AI may be effectively utilized within its regulatory architecture. It then examines the relationship between medical AI and sector-specific healthcare legislation, particularly the Digital Medical Products Act and the Medical Devices Act, focusing on the regulation of AI-enabled medical devices and Software as a Medical Device (SaMD). The Article further explores how the distinctive characte- ristics of medical AI—such as ensuring safety and effectiveness, managing algorithmic changes, addressing explainability, and securing cybersecurity—pose significant challenges to traditional medical device regulatory frameworks. Special attention is given to liability-related issues arising from the clinical use of medical AI, including the duty of care owed by health care professionals and evidentiary difficulties stemming from the black-box nature of artificial intelligence. The Article argues that, under the current legal framework, the AI Framework Act is likely to function primarily as a declaratory and principle-oriented statute, rather than as a self-executing regulatory instrument, and that effective governance of medical AI therefore requires systematic coordination with sector-specific legislation. In conclusion, this Article contends that the regulation of medical AI should not rely on a single, comprehensive statute. Instead, the AI Frame- work Act should operate as a general principles statute, while health care- specific laws—most notably the Digital Medical Products Act and the Medical Devices Act—should provide concrete and enforceable regulatory standards. This layered regulatory approach offers an interpretive frame-work for ensuring the safe deployment of medical AI and for clarifying the allocation of legal responsibility in the evolving landscape of AI-driven health care.