인공지능과 법인격 - 불법행위책임의 관점에서 -

The rapid development of artificial intelligence(hereinafter ‘AI’) is raising various issues in civil law, such as legal personhood, legal acts, validity of contract, and the subject of liability. In particular, in terms of tort liability, the subject of responsibility, the basis for responsibility, and the content of responsibility in the ‘AI accident’ are placed on the main agendas. The issue that should be discussed in relation to the ‘subject of responsibility’ is whether the AI ​​itself that caused the accident can be held responsible, that is, whether the AI ​​can be recognized as a legal person. This paper summarizes the discussions regarding the issue of recognition of AI legal personhood raised in various fields such as natural science, engineering, psychology, and philosophy as well as law. After reviewing, conclusions were presented. From a scientific and technological point of view, AI cannot be recognized for its ‘autonomy’ or the same level of behavior as human beings. From a philosophical point of view, it is difficult to see that legal positivism or post-anthropocentrism provides a firm rationale for overcoming the ideas of natural law and humanism and affirming the legal personhood of AI. In terms of legal strategy and policy, it is difficult to recognize the social reality or legal convenience that should affirm the legal personhood of AI, and there is a substantial risk of causing considerable social costs and legal confusion. Therefore this paper concludes that AI should not be recognized as legal person from the perspective of tort liability. In short, the AI itself that caused the accident cannot be held liable for tort, and it should be discussed on ‘who’ should be held responsible ‘for what’ and ‘on what’ basis in case of AI accident.