Text and data mining (TDM) is crucial for generative AI models to create content. However, due to its nature of reproducing large amounts of data and copyrighted works, TDM often finds itself in a challenging relationship with copyright law, which traditionally protects copyrighted works. While there has been significant discussion and research on TDM exceptions and limitations under the EU Digital Single Market (DSM) Directive and other national laws, South Korea's approach to TDM legislation, particularly for generative AI, has received comparatively less attention. Recognizing the importance of TDM, there is broad agreement on the need to include specific copyright exceptions for TDM in the Korean Copyright Act. This article aims to introduce and analyse the proposed amendment to the Korean Copyright Act concerning TDM exceptions. In doing so, it will present and compare the latest legislative developments in Korea with the global trends in TDM exceptions.
Keywords: generative AI; text data mining; TDM; exceptions and limitations; DSM Directive; non-enjoyment