摘要
深度伪造作为人工智能深度学习应用的具体场景,当前已带来严重的社会风险,尤其是针对公众人物的深度伪造,更可能对社会秩序和国家安全造成严重冲击与损害,因而引起了世界各国的高度重视。深度伪造的技术本质是人工智能的算法滥用,行为本质是个人生物识别信息的滥用。美国和欧盟等分别采取了专门立法以及借助现有的"数据-信息"规制路径来防范深度伪造的社会危害,在法益定位、立法重点和法律模式上均存在差异。我国刑法目前基于目的性行为的结果归责思路和基于公民个人信息保护的前端归责思路忽略了深度伪造法益侵害的独立性以及个人生物识别信息保护的特殊需求,无法完全实现刑法的规范目的。深度伪造的规范本质是身份盗窃行为,有必要在刑法中引入身份盗窃,既能建立"公民个人信息-身份信息-生物识别信息"的梯次加重保护体系,弥补"合法获取+不法利用"个人信息的刑法评价空白,并顺带规制传统的身份盗窃行为,增强身份盗窃入罪化的扩散性立法效应。
As a specific pattern of artificial intelligence deep learning application,"deepfake"has brought serious social risks,especially the"deepfake"against public figures,which is more likely to cause serious impact and damage to social order and national security.Therefore,it has attracted the attention of all countries in the world.The technical essence of"deepfake"is the abuse of artificial intelligence algorithms,and the behavioral essence is the abuse of biometric information.At present,some problems exist both in the result attribution based on"purposive behavior"and the front-end attribution based on protection of"citizen personal information"."Identity theft"behavior is the core for us to regulate"deepfake".It is necessary to introduce"identity theft"in the criminal law,which can not only establish a stepby-step protection system for"citizen personal information-identity information-biometric information",but also make up for the criminal law evaluation blank of"legal access&illegal use"of personal information.In this way,we can also achieve effective regulation of traditional identity theft and diffusive legislative effect of criminalization of identity theft.
出处
《政法论坛》
CSSCI
北大核心
2020年第4期144-154,共11页
Tribune of Political Science and Law
基金
国家社科基金青年项目“数据开放的刑法边界研究”(17CFX022)的阶段性研究成果
中国政法大学青年教师学术创新团队支持计划资助。
