生成式人工智能背景下个人信息的刑法分类研究  

A Study on the Criminal Law Classification of Personal Information in the Context of Generative Artificial Intelligence

作　　者：李冲[1] LI Chong(Shandong University,Qingdao,266237,Shandong)

机构地区：[1]山东大学,山东青岛266237

出　　处：《湖南警察学院学报》2023年第4期40-47,共8页Journal of Hunan Police Academy

基　　金：2022年中国博士后科学基金第72批面上资助“少捕慎诉慎押政策背景下羁押听证制度研究”(2022M721918);2022年山东省高等学校“青创团队计划”团队“(刑事合规立法问题创新团队)刑事合规制度的立法方案研究”(2022RW004)的阶段性成果。

摘　　要：数据分类是党中央、国务院确定的化解数据风险的基础手段。生成式人工智能背景下,个人信息“基于权利的分类方法”存在界限模糊、权力不对称、阻碍技术发展的问题。相比之下,个人信息“基于风险的分类方法”有利于明确分类范围、平衡风险责任、促进信息流通。在个人信息“基于风险的分类方法”中,“风险”体现为法益受侵犯的程度和概率,可以在社会情景、处理主体、信息属性和传输原则组成的具体场景中进行确认。在实践中,刑法可以利用高风险、中风险、低风险三分法对个人信息进行功能性分类,通过场景化路径,借鉴已有司法案例来实现对信息处理主体的罪责判定。Data classification is the fundamental means for resolving data risks determined by the Central Committee of the Communist Party of China and the State Council.In the context of generative artificial intelligence,the"rights-based classification method"of personal information has problems such as blurred boundaries,power asymmetry,and obstacles to technological development.In contrast,the"risk-based classification method"for personal information is conducive to clarifying the scope of classification,balancing risk responsibilities,and promoting information circulation.In the"risk-based classification method"of personal information,"risk"reflects the degree and probability of infringement of legal interests,which can be confirmed in specific scenarios composed of social scenarios,processing subjects,information attributes,and transmission principles.In practice,criminal law can use the high-risk,medium risk,and low-risk triad to classify personal information functionally.Through a scenario based approach,existing judicial cases can be used to determine the criminal responsibility of information processing subjects.

关 键 词：个人信息 生成式人工智能 刑法分类 

分 类 号：D924.3[政治法律—刑法学]