A Study on Copyright Ownership of AI-Generated Content: A Case Study of ChatGPT and DeepSeek

Abstract

The widespread application of generative AI technology has led to a global trend in disputes over content ownership. This study selects two representative generative models, ChatGPT and DeepSeek, as its research subjects. Through case analysis, text deconstruction, and comparative research, it systematically examines 23 typical domestic and international judicial precedents and 12 platform agreements from 2021 to 2024. The empirical analysis reveals three core issues: First, the degree of creativity (specificity and structure) and the amount of content modification in user prompts are key factors in copyright determination; second, the ownership clauses unilaterally drafted by platforms suffer from formatting defects, with 78% of these clauses suspected of violating Article 497 of the Civil Code's prohibition on "excluding users' primary rights"; and third, the current Copyright Law limits creators to natural persons, resulting in an ownership vacuum for algorithm-generated content. Based on this, we propose a three-dimensional system: establishing a quantitative contribution evaluation system encompassing "instruction-generation-modification," adding a special registration system for algorithm-generated content, and promoting adaptive reforms in copyright law regarding the scope of rights holders, originality recognition standards, and the boundaries of platform responsibility. This study provides theoretical support and practical references for developing new copyright rules for the AI era.