論人工智能生成內容在著作權法中的界定

On the Copyrightability of Artificial Intelligence Generated Contents in Copyright Law

人工智能生成內容是否構成著作權法下的作品，以及如果構成作品，著作權應歸何方所有，是當前人工智能應用中的一個有待明確的問題。當前關於人工智能生成內容是否構成作品的觀點可大體從「獨創性」要件的不同理解分為主體視角、行為視角和客體視角三類。在保障人類創作空間、維繫人類尊嚴和文化繁榮，同時為人工智能內容生成產業提供合理的激勵，以推動其良性發展的多重目標之下，可按客觀視角從作品形式上認定其獨創性。對於著作權歸屬，應在約定優先的前提下，默認著作權歸人工智能系統使用者所有，以明確人工智能的工具地位，促進在人工智能作品基礎上的人類創作。

One of the key issues in the current copyright law is whether the artificial intelligence generated content constitutes a work under the copyright law and whom the copyright of the work should belong to. The current main points of views can be divided into three categories: the subjective perspective, the behavioral perspective and the object perspective. To leave enough creative space to human beings, maintain human dignity and cultural prosperity, and to provide reasonable incentive for the artificial intelligence content generation industry, the originality of artificial intelligence generated content can be determined in the “object perspective”. The copyright should belong to the user of the artificial intelligence system by default, thus to clarify that artificial intelligence should be viewed as a tool, and promote human creation based on artificial intelligence.