人工智能生成物可版权性的法律判断方法辨析  

Analysis of Legal Judgment Methods for the Copyrightability of Artificial Intelligence Product

作　　者：魏晓东[1] Wei Xiaodong

机构地区：[1]辽宁大学法学院、辽宁知识产权学院

出　　处：《盛京法律评论》2023年第1期82-107,共26页Shengjing Law Review

摘　　要：人工智能技术的利用不应成为其生成物不予版权保护的法律上的理由。相反,人工智能技术能减少社会的无谓损失,激励人类创作,提高文化、科技作品创作水平,具有增进社会总体福利的功能。通过甄别人工智能技术参与与否来决定是否施加版权保护是无效的,背离了《著作权法》的基本价值目标,也违反了人工智能技术发展规律。具备创造性目标功能的人工智能技术以模拟人类大脑为终极目标,其采取的无监督学习算法技术识别起来成本极高,对这样技术的生成物进行甄别而不予版权保护,会损害法律追求的基本效率价值。Artifieial intelligence technology should not be an excuse to deny copyright protection to its products.On the contrary,artificial intelligence technology has the functions of reducing social deadweight loss,stimulating human creation,improving the creation level of works,and enhancing the overall social welfare.It is invalid to determine whether copyright protection is created by screening AI existence,which deviates from the basic value goal of copyright law and violates the development rule of AI technology.The ultimate goal of artificial intelligence technology with creative goal function is to simulate human brain.The unsupervised learning algorithm technology adopted by artificial intelligence technology costs a lot to identify the effect,and the works generated by such technology will not be protected by copyright after screening,which will damage the basic efficiency value of law.

关 键 词：人工智能 作品 版权保护 著作权法 创作成本 

分 类 号：D923.41[政治法律—民商法学]