생성형 인공지능 학습을 위한 저작권 라이선스 모델의 제도적 분석: 권리 보호와 기술 발전의 균형 모색

The rapid advancement of generative artificial intelligence technology has triggered legal debates over the legitimacy of training data and the ownership of generated outputs. This study begins by defining key AI-related concepts and then takes the AI licensing model of GEMA, the German music copyright management organization, as a point of departure. By incorporating major cases such as the lawsuit filed by GEMA against Suno AI, the paper conducts an in-depth examination of the institutional logic, operational mechanisms, and implementation challenges of various copyright licensing models, offering insights for future copyright governance. Copyright-protected works used for AI training are increasingly shifting toward licensed usage, with both collective and individual licensing models gaining traction. The study assesses copyright licensing as a viable and scalable approach to aligning technological progress with rights protection. Furthermore, it argues that establishing an AI copyright regime that balances incentives for innovation with the safeguarding of rights requires an interplay between international coordination and the creative development of domestic legal systems, thereby enabling the sustainable coexistence of technological advancement and the creative industries.