Determining the Infringing Subject and Allocating Liability in Copyright Violarion of AI-generated Content

With the rapid advancement of artificial intelligence (AI) technology, incidents of copyright infringement involving AI-generated content have become increasingly prevalent. Against this backdrop, identifying the infringing subject in generative AI-related copyright cases and establishing rules for liability allocation have become urgent legal issues. This paper first analyzes the copyright attributes of AI-generated content, pointing out the difficulties in applying traditional criteria for determining infringing subjects and the controversies surrounding the application of the safe harbor rule and red flag standard. It then explores four potential infringing subjects—AI itself, AI developers, AI service providers, and AI users—by examining their roles and responsibilities in infringement scenarios. Through this analysis, the paper confirms the primary liability of AI service providers as infringing subjects. This paper examines how liability should be assigned to AI service providers in copyright infringement cases. It argues that the traditional fault-based liability principle is difficult to apply effectively in this context due to practical and evidentiary challenges, and that the presumed-fault principle also suffers from significant theoretical and procedural shortcomings. It concludes that adopting a strict liability (no-fault) principle for AI service providers is a more appropriate approach. However, practical measures must be taken to reasonably limit and improve this liability, including limitation rules, liability insurance mechanisms, and compensation funds. Using a benefit-balancing method, this paper seeks to harmonize the protection of creators' rights with technological innovation, aiming to unify legal practice in AI-related copyright infringement cases.