AI copyright disputes intensify in US and South Korea

Legal and regulatory debates regarding artificial intelligence and copyright are intensifying in both the United States and South Korea.

In the United States, a legal battle is unfolding between media organizations and AI companies. The U.S. Department of Justice submitted an amicus brief to a New York federal court arguing that using copyrighted text to train Large Language Models (LLMs) constitutes ‘transformative’ fair use. The Department distinguishes between the training phase, which digitizes knowledge, and the output phase, which may reproduce specific works. However, news outlets like The New York Times argue that this stance prioritizes AI industry competitiveness over the rights and compensation of content creators, warning that AI summaries could replace the need to visit original news sites.

In South Korea, discussions are shifting toward establishing practical industry rules for AI-generated music. A recent National Assembly forum addressed how to determine human creative contribution and how to verify AI usage in music licensing. Experts are debating whether the level of control a musician exerts over an AI tool constitutes authorship, or if simply selecting an AI-generated phrase is merely ‘discovery’ rather than ‘creation.’ There are ongoing calls to include the voices of active composers in these regulatory discussions to ensure the resulting systems support rather than undermine professional creators.