Battle between ‘Long’ and ‘Short’ Videos: Fragmented Uses, Diversified Purposes, and the Evolution of China’s Copyright Limitation Rules in the AI Era

Abstract The rapid proliferation of artificial intelligence (AI) technology, particularly the deployment of generative AI in the creation of short videos, has exposed the anachronistic and incomplete nature of the copyright limitations and exceptions (L&E) within the Copyright Law of China. The US jurisprudential approach, characterized by the four-factor fair use analysis and the doctrine of transformative use, has been embraced in China by courts as ‘general legal principles’. These principles have been localized within the restrictive framework of the Berne Convention’s three-step test. However, the application of these principles to China’s burgeoning AI and AI-generated content industry, exemplified by the short video sector, reveals a complex landscape. The purposes behind the utilization of copyrighted works are multifaceted, and the intense rivalry between ‘long’ and ‘short’ videos that represent established and emerging giant Internet platform enterprises further muddies the waters in the application of the American ‘purpose transformation’ concept. A better alternative would be to seek guidance from the Copyright Law of Japan. By grounding decisions in pre-existing, enumerated copyright exceptions, systematically categorizing L&E scenarios, and delineating key considerations for various categories, the judiciary can be steered towards a more nuanced and equitable resolution of individual cases and strike a balance between the interest of copyright holders, Internet users, and the emerging Internet platform enterprises.