Research on the Criminal Law Protection of Copyright for Artificial Intelligence Creations

The rapid development of generative artificial intelligence (AI) has led to the widespread production and commercial use of AI-generated content (AIGC), raising new questions concerning copyrightability, rights ownership, and criminal regulation of infringement. From the perspective of coordination between civil and criminal law, this article examines the copyrightability and ownership of AIGC and the corresponding boundaries of criminal-law protection. It argues that the involvement of AI should not, by itself, exclude generated content from copyright protection. The decisive inquiry should be whether a human user has made sufficient creative contributions through prompt design, selection, modification, and refinement of outputs. Where such contributions satisfy the requirements of copyright law, the user may in principle enjoy copyright in the resulting work. Criminal-law protection should then be premised on rights recognized under copyright law and constrained by the principles of legality and restraint. AIGC-related infringement can be divided into two principal categories: infringement of copyrightable AIGC and infringement of pre-existing works through generative AI. For the latter, model training and user-side use should be distinguished, and unauthorized input should not automatically establish infringement at the output stage. Criminal liability should depend on whether protected expression is substantially reproduced and whether the statutory act, quantitative threshold, and subjective intent required by Article 217 of the Criminal Law are satisfied.