论人工智能生成内容的不可版权性

On the Non-Copyrightability of Content Generated by Artificial Intelligence

摘要: 随着人工智能生成内容的大量出现，其生成内容是否具备可版权性成为研究应对人工智能生成内容的核心问题之一。作者中心主义相对于读者中心主义更为合理，独创性的认定必须同时具备主观标准和客观标准，两者不可分离。人工智能不可拟制为“机器作者”，从而使其具备主观标准。根据“普全数理”能够赋予人工智能以人的意志，但现阶段人工智能仍然无法实现这一理论。人工智能亦无法通过工具论借助人类自由意志使其生成内容成为作品。是否符合立法目的是法律规则是否正当的基本判断标准之一。人工智能生成内容进行版权化与著作权法立法目的南辕北辙，甚至解构著作权法的存在基础。

Abstract: With the emergence of a large amount of content generated by artificial intelligence, whether the generated content has copyrightability has become one of the core issues in researching and dealing with AI generated content. Author centrism is more reasonable than reader centrism, and the recognition of originality must have both subjective and objective standards, which cannot be separated. Artificial intelligence cannot be portrayed as a “machine author”, thus possessing subjective standards. According to “universal mathematics”, artificial intelligence can be endowed with human will, but at present, artificial intelligence still cannot achieve this theory. Artificial intelligence is also unable to generate content into works through instrumental theory and human free will. Whether it meets the legislative purpose is one of the basic criteria for judging the legitimacy of legal rules. The copyrightization of content generated by artificial intelligence goes against the legislative purpose of copyright law, and even deconstructs the foundation of copyright law.