의료 인공지능 자율성에 따른 형사책임 귀속 연구

This paper raises the need for a criminal law response to this on the premise that conflicts with existing criminal law theories are occurring due to the strengthening of autonomy of medical artificial intelligence. In particular, medical artificial intelligence is not only directly involved in human life, but also has a high risk in that negligence due to a single error can damage many patients. Due to this specificity, it is difficult to fully explain or respond to the problem situation caused by medical artificial intelligence only with the human-centered responsibility theory of the existing criminal law. Therefore, according to the level of technology autonomy, this paper classified medical artificial intelligence into ① simple medical aid type, ② quasi-autonomous type, and ③ fully autonomous type, and reconstructed the scope and legal structure of responsibility for each stage. By combining the “autonomy” element with the existing risk-oriented approach, the degree and standards of responsibility that can be imposed on related actors such as medical personnel, designers, and manufacturers were tried to be more elaborate. In the case of fully autonomous artificial intelligence, the applicability of indirect criminal theory based on social behavior theory was also examined. This is an alternative attempt to prepare for the emergence of advanced autonomous artificial intelligence, which is difficult to embrace artificial intelligence only with the concept of “tools” under the existing criminal law, and for the flexible reconstruction of the criminal responsibility structure. As a result, this paper seeks to establish a normative framework to prevent technological development from colliding with human rights guarantees, but to have the ultimate purpose of research, it seeks to harmonize the dual tasks of criminal law theory and protection of basic rights through criminal responsibility structures that reflect the level of autonomy of medical artificial intelligence.