인공지능(AI) 관련 규범과 연성법의 역할에 관한 소고 - 디지털경제협정을 중심으로

This paper aims to analyze the role and legal and policy implications of soft law in the establishment of hard law provisions in the Digital Economy Agreement in the process of creating international norms for artificial intelligence (AI) technology. The AI field is an area where the pace of technological development is fast, has a wide-ranging social and economic impact, and where the economic interests of each country are intricately intertwined with the rule-making of the field. Therefore, the traditional form of international law, so called the hard law approach, has the limitation that it is not easy to create timely and effective rules. As a result, soft laws on ethical principles, technical standards, and policy guidelines agreed in the field of AI governance, led by international organizations such as the OECD, UNESCO, and the UN, and multilateral forums such as the G7, have played a leading role in shaping rules. First, this paper aims to confirm that the AI provisions included in the Digital Economy Agreement, which is a form of hard law, are mostly composed of non-binding expressions such as consideration of international principles when developing an ethical governance framework and cooperation to ensure technological interoperability, which is a process in which soft law is reflected in hard law in a way that institutionalizes the system. Second, we will examine the phenomenon of ‘hardening of norms’ in which soft law is gradually institutionalized and converted into hard law through examples of rule-setting such as the e-ASEAN Framework Agreement and the promotion of DEFA. Third, the countries such as the UK, Singapore, and South Korea are taking the lead in shaping AI rules within the Digital Economy Agreement, while the U.S. and China are relatively passive. The EU, on the other hand, cites AI related provisions as the basis for its regulatory approach to justify cross-border data transfer provisions, highlighting the differences in national regulatory strategies. Through this analysis, the paper finds that the AI provisions in the Digital Economy Agreement are being constructed in accordance with the convergent operation of international rules based on soft law, which functions as a hybrid normative structure of soft and hard law that seeks to ensure the consistency and reliability of AI regulations between countries. In other words, soft law is not just a transitional measure for the regulation of emerging technologies such as AI, but is establishing itself as a sustainable tool in the process of shaping international law.