人工智能创造物不真实署名的风险与规制  

Research on the Risk and Regulation of Untrue Signature of Artificial Intelligence Creation

作　　者：姚志伟 沈燚 YAO Zhiwei;SHEN Yi(Research Center on the Law and Economic Development,Guangdong University of Finance and Economics,Guangzhou 510320,China;Research Center on the Internet and Big Data Law,Shanghai University of Finance and Economics,Shanghai 200433,China)

机构地区：[1]广东财经大学法治与经济发展研究所,广东广州510320 [2]上海财经大学法学院网络与大数据法治研究中心,上海200433

出　　处：《西安交通大学学报（社会科学版）》2020年第1期133-140,共8页Journal of Xi'an Jiaotong University：Social Sciences

基　　金：国家社会科学基金项目(18VHJ011)

摘　　要：在现行法下,人工智能创造物的著作权保护面临作品认定难题和权属分配难题。解决这两个难题存在两条思路:一是改革现行著作权法,为人工智能创造物设计新的规则;二是不对著作权法进行特别的改革,而从实然层面解决问题,即通过不真实署名,将人工智能创造物伪装成人的作品,从而获得著作权法的保护。第二条思路避开了重新设计规则的难题,但是在实践中会面临困难。同时,署名行为的影响具有复合性,其影响不限于著作权范围。不真实署名会冲击以作品为基础的社会评价机制,且存在公共安全风险。2019年5月公布的《数据安全管理办法(征求意见稿)》对不真实署名行为进行了规制,基本的规制方法是披露,即披露信息由人工智能所创造。该规则尚有待完善之处,包括披露的方式应改进、披露的范围应缩减、披露义务的承担主体应更明确。Under the current legal system, the copyright protection of artificial intelligence creation has faced the difficulty of the identification of the works and the assignment of the ownership. There are mainly two ways to solve this problem: one is to reform the current copyright law, to design new rules for artificial intelligence creation;the other is to solve problems from the actual level instead of carrying out special reforms on the copyright law, which disguised the artificial intelligence creation as a human work by untrue signature, thereby obtaining the protection of the copyright law. It can clearly be seen that the second way avoids the difficulty of redesigning the rules, but it will face some difficulties in real practice. At the same time, the legal consequences of the signature act are compound, and the impact is not limited to the copyright law. Obviously, untrue signatures can impact work-based social evaluation mechanisms and have public safety risks. Moreover, "Data Security Management Measures(Draft for Comment)", which was promulgated in May 2019, regulated untrue signature behaviors by basic regulatory method, which included the disclosure of the fact that the information is created by artificial intelligence. However, the rules still need to be improved, including the way in which disclosure should be improved, the scope of disclosure should be reduced, and the subject of disclosure obligations should be more explicit.

关 键 词：著作权保护 人工智能创造物 不真实署名 披露 社会评价机制 公共安全风险 

分 类 号：D922.16[政治法律—宪法学与行政法学] TP18[政治法律—法学]