인공지능의 콘텐츠 창작에서 저작물 이용의 법적 쟁점에 관한 고찰

Artificial intelligence(AI) learns existing copyrighted works through machine learning like human beings and uses them with a rapid speed in the process of creation of new contents. The purpose of this study is to examine how the use of works or data by AI inevitable to learning process should be evaluated under the current copyright regime and how it should be discussed from the point of view of copyright infringement and fair use. For this purpose, the study conducts qualitative depth interviews by e-mail with experts such as scholars in law and media and researchers on policy, legislation, and ICT, and analyzes the views of interviewees and their implications. The research results show that the interviewees significantly consider factors such as the function and technology of AI, the purpose and characteristic of the use of works, and the possibility of human control in the process of learning in discussing the meaning and the legal character of the use of works. In addition, the interviewees concur with the view that there is no difference in defining the use of works by human beings and that by AI, on the basis that AI is only used as a creation tool of human beings and the ultimate purpose of use is creation. They are also positive to the idea that licensing system should be differentiated from that of human beings with method and procedure specific to the use of works by AI in the process of learning. It is founded that the interviewees think that the types of potential copyright infringement in the learning process are infringements of the right to copy, transmit, and produce derivative works and the right to the integrity. It is also shown that they think that liability for the copyright infringement in the use of work by AI should be imposed on human beings. The result shows that the purpose of use of works works as an important factor in the judgment of fair use of intermediate copying in the process of learning by AI. When it comes to technological fair use, they are negative on the basis that technology advancement should not be used as the legal basis for justifying fair use and it is against the basic spirit of fair use and cultural value creation. Based on the results, this study suggests introducing new copyright exception for data mining and special provision on fair use to protect learning by AI and building up the system for providing data for learning by AI in the public sphere and the international copyright guideline for the creation of new works by AI.