생성형 인공지능의 저작권 침해 예외 규정으로서 공정이용 법리의 유효성에 관한 시론적 고찰

Abstract

As generative artificial intelligence is commercialized throughout society as ‘generative AI’, the debate over intellectual property rights and copyright protection is entering a new phase. Unlike existing artificial intelligence, generative artificial intelligence allows users to directly issue commands using human language rather than coding, and recognizes visual data such as pictures, photos, and videos directly through a variety of multimodal programs being developed. It is possible to express it in human language in the interface. In addition, unlike existing artificial intelligence that utilizes expert knowledge in a specific field, it has the greatest differentiation in that it functions as a ‘general artificial intelligence’ (generative AI) that can be utilized comprehensively throughout society. The problem of copyright infringement in generative artificial intelligence is being discussed from two aspects. First, the issue of copyright infringement on the vast amount of data learned by generative artificial intelligence during the development stage, and second, the issue of copyright infringement on the vast amount of data learned by generative artificial intelligence during the development stage. This is an issue of copyright infringement by various products. Of course, existing artificial intelligence models also used data for learning, and there have been conflicts over copyright infringement as a result. And from the perspective of promoting the artificial intelligence industry, the law has been considering various methods, such as the legal principles of fair use, TDM exemption provisions, and terms of use, to apply exceptions to copyright infringement. However, with the emergence of generative artificial intelligence, various laws that have been discussed for exceptions to copyright infringement are entering a new phase. The TDM exemption regulations that countries around the world had rushed to announce were postponed, and in the case of the United States, the strength of fair use review, which has been maintained to recognize exceptions to copyright infringement since the Campbell precedent, has been strengthened in recent precedents. Meanwhile, in our case, the TDM exemption regulation was announced as included in the “Copyright Act Amendment” at the end of 2021, similar to foreign cases, but was postponed with the emergence of generative artificial intelligence in 2022. However, in our recent case law applying the legal principle of fair use, a ruling was made similar to the fair use judgment criteria in the US case law before the development of generative artificial intelligence. So, is it necessary to change our precedents in the era of generative artificial intelligence, or should we grant comprehensive rights to create secondary works by broadly citing transformative use and promote new industrial fields through this? Discussion will be needed. Accordingly, in this paper, we analyze the legal principle of fair use that has been applied as an exception to copyright infringement in Korea and the United States and the changes in the fair use legal principle in recent U.S. cases, and the legal principle of fair use is used to apply generative artificial intelligence as general artificial intelligence. Through a theoretical consideration of whether it can still be applied effectively even in the new environment of its emergence, we would like to discuss that the legal principle of fair use still has some validity in various copyright infringement cases, such as data learning of generative artificial intelligence.