Determining the Rights and Attributes of Generative AI- Generated Objects —Based on the Civil Code's "Civil Rights and Interests" System

Abstract

The widespread application of generative artificial intelligence technology has made the determination of the rights attributes of its generated works a core legal challenge that urgently needs to be addressed. Current mainstream discussions in academia and practice are confined to the framework of copyright law, focusing on whether generated works constitute "works" and the source of their "originality." However, this has reached a stalemate due to the difficulty in overcoming the theoretical consensus on "human author-centrism" and the protection of "intellectual achievements." To break through the constraints of the existing paradigm of copyright law, this article advocates shifting the focus to the "civil rights" system established by the Civil Code of the People's Republic of China. This system does not necessarily take direct human intellectual creation as the sole premise, and can effectively circumvent the inherent obstacles to the protection of human intellectual achievements under copyright law. This article divides the rights attributes corresponding to generated works into three different levels: purely AI-generated, limited human participation, and deep human participation. This tiered determination system not only aligns with the technological nature of human-machine collaboration in generative AI, but also provides clear and operable core discretionary benchmarks for judicial decisions, thereby resolving the dilemma of unclear ownership and ambiguous rights allocation in practice, and providing theoretical reference and practical guidance for including new types of objects in the digital age within the scope of civil rights protection.