表达型人工智能作品的侵权分析——以复制和文本数据挖掘为对象  

Infringement Analysis of Expressive AI Works:Targeting Copying and Text Data Mining

作　　者：马驰升 黄霜瑜 MA Chisheng;HUANG Shuangyu(School of Law and Public Administration,Hunan University of Science and Technology,Xiangtan,Hunan 411201,China)

机构地区：[1]湖南科技大学法学与公共管理学院,湖南湘潭411201

出　　处：《闽江学院学报》2024年第4期70-78,共9页Journal of Minjiang University

基　　金：湖南省哲学社会科学基金青年项目(21YBQ058);湖南省普通本科高校教学改革研究项目(202401000925)。

摘　　要：人工智能作品的生成过程包含了复制和文本数据挖掘两种可能涉嫌侵犯版权的行为,二者相对应的抗辩理由分别是“合理使用原则”和“二分法”。通过对人工智能作品的生成过程和分类的研究,认为判断复制行为是否适用合理使用原则来排除侵权的要点在于“人工智能作品是否会对被学习作品版权人的利益产生影响”,判断文本数据挖掘行为是否侵权的要点在于“人工智能是否对原作的表达进行了利用并外化为其作品内容”。厘清这两个要点有助于识别人工智能作品产生过程中的侵权行为,同时也有利于促进数字时代对版权的保护。The generation process of AI works includes"copying"and"text data mining",which are probably suspected of infringing the copyright of these learned materials.The corresponding defenses are"Principle of fair use"and"dichotomy".By studying the generation process of AI works and the classification of their works,we can draw two conclusions:The key point of judging whether the copying can apply the principle of fair use to exclude infringement lies in"whether the artificial intelligence works will affect the interests of the copyright owners of the studied works",and the key point of judging whether text data mining will infringe the copyright lies in"the degree of correlation between the content of artificial intelligence works and the copyright protected part of the learned works".These two conclusions are helpful to identify the infringement in the production of artificial intelligence works,and also to promote the protection of copyright in the digital age.

关 键 词：人工智能作品 复制 文本数据挖掘 合理使用 表达性使用 非表达性使用 

分 类 号：D923.4[政治法律—民商法学]