人工智能生成内容的版权问题研究

A Study of Copyright Issues in Artificial Intelligence Generated Content

摘要: 创作型人工智能ChatGPT的横空出世，使得人工智能生成内容的版权问题愈加复杂。人工智能是否具有法律主体资格和其所创作的内容是否构成作品是首先需要回答的问题。由于缺乏独立意志，使得其仍然难以拥有法律主体资格。人工智能生成内容根据客观标准和主观标准，对于独创性的判断截然不同。人工智能在机器创作中可能引发的侵权风险是最亟待解决的版权冲突。对于数据输入行为，各国存在“无条件例外”与“有条件例外”等不同做法，纳入合理使用或是现在最适合我国的路径。对于其所生成内容的版权侵权，目前仍需采取“接触 + 实质性相似”，以过错原则为主进行判断。

Abstract: The emergence of ChatGPT, a creative AI, has made the issue of copyright for AI-generated content increasingly complex. Whether AI has legal subject status and whether the content it creates constitutes a work are the first questions to be answered. The lack of independence will still make it difficult for it to have legal subject status. AI-generated content is judged very differently for originality based on objective and subjective criteria. The risk of infringement that AI may cause in machine creation is the most pressing copyright conflict to be resolved. In terms of data input, there are different approaches such as “unconditional exception” and “conditional exception” in different countries, and the inclusion of fair use is the most suitable path for our country now. As for the copyright infringement of the content generated by AI, it is still necessary to adopt the principle of “contact + substantial similarity” and the principle of fault as the main judgement.