摘要
【目的】对人工智能生成物的法律属性进行准确定位,寻找人工智能生成物的潜在风险化解路径。【方法】从法律层面分析人工智能生成物的产生来源及过程,揭示人工智能生成物的本质。【结结果果】从人工智能“数据训练”创作方式和机器学习阶段分析来看,人工智能不具备民事主体资格,但其作品具有可版权性的可能性,应被赋予一定权利。【结论】赋予人工智能享有署名权的权利,以“使人工智能生成物成型确定下来”为分界线,在自然人范围内确定权属和责任承担者进行权利安排。运用实质性相似规则和市场替代标准化解人工智能生成物复制权和改编权侵权风险,同时适用过错原则,相应配置人工智能设计者、使用者“防抄袭设计义务”注意义务。
[Purposes]This paper aims to accurately locate the legal attributes of artificial intelligence products and find the potential risk resolution path of artificial intelligence products.[Methods]The nature of artificial intelligence products is revealed through legal analysis of the source and process of artificial intelligence products.[Findings]From the analysis of the creation method of artificial intelligence"data training"and the machine learning stage,artificial intelligence does not have the qualification of civil subject,but the works have the possibility of copyright,which should be granted certain rights.[Conclusions]This paper gives artificial intelligence the right to enjoy the right of signature,and with"the formation of artificial intelligence products to be determined"as the dividing line,this paper determines the ownership and responsibility bearers within the scope of natural persons to arrange the rights.The substantive similarity rules and market substitution standards are used to resolve the risk of infringement of reproduction rights and adaptation rights of artificial intelligence products.At the same time,the principle of fault is applied,and the duty of care of"anti-plagiarism design obligations"of artificial intelligence designers and users is configured accordingly.
作者
东村仁
臧可
DONG Cunren;ZANG Ke(College of Intellectual Property,Henan University of Economics and Law,Zhengzhou 450046,China;School of Law,Northwest Normal University,Lanzhou 730070,China)
出处
《河南科技》
2024年第6期116-123,共8页
Henan Science and Technology
关键词
人工智能生成物
独创性
实质性相似
过错推定
artificial intelligence products
originality
substantial similarity
presumption of fault