Standards for Copyrightability of AI-Generated Content (AIGC): Theoretical and Practical Perspectives

Abstract

The explosive development of generative artificial intelligence technologies, exemplified by ChatGPT, has fundamentally challenged traditional Copyright Law by introducing new models of human-machine collaborative creation, making the copyrightability of AI-generated content (AIGC) a cutting-edge issue in contemporary judicial practice and academic studies. This technological wave has directly given rise to several typical cases in China, where courts have grappled with unprecedented legal questions. The current academic debate encompasses three main viewpoints: the negative view emphasizes that copyrightable works must be originated directly from natural persons and argues that artificial intelligence fundamentally lacks authorship capacity; the affirmative view regards AI as merely a sophisticated tool or active co-creator with human users; while the neutral view advocates for case-by-case analysis, focusing particularly on examining the causal relationship between user prompts and the final generated contents. The core controversies involve applying the traditional thought/expression dichotomy, identifying the extent of user control over outputs, and coordinating international protection standards, etc. Against this background, this paper aims to shift the identification framework of AIGC copyrightability from a conventional result-oriented approach to a more nuanced process-oriented one, proposing specific and operational standards that can help unify judicial standards, provide practical references for courts, and offer a solid doctrinal basis for the adaptive interpretation of Copyright Law in the artificial intelligence era. Establishing appropriate criteria is crucial for promoting cultural prosperity and innovation without either stifling legitimate AI-assisted creation or overprotecting low-quality, machine-dominated outputs.