摘要
输入阶段获取的训练语料是生成式人工智能构建数据库的基础,现行刑法尚不足以抵御此阶段带来的个人信息保护风险。生成式人工智能获取海量信息的需求与刑法的控制保护理念天然产生冲突,未经权利主体同意收集个人信息的行为不仅影响了公民个人信息自决权的行使,还带来了隐私泄露的风险,冲击着个人信息的安全与管理秩序。传统个人信息保护模式下的规定不足问题日益严峻,公开个人信息的处理与侵犯公民个人信息罪司法解释中“情节严重”判断陷入困境。因此,应从三方面着手提升新形势下刑法对个人信息保护的应对能力:首先,将公开个人信息分为主动公开、被动公开和非法公开的三类进行保护;其次,在刑法领域,收集个人信息的行为需要遵循合法、正当和必要的原则;再次,将个人信息类型作为“情节严重”判断的第一标准并完善个人信息的类型。
The training data obtained in the input stage serves as the foundation for the database construction of generative ArtificialIntelligence(AI)The current criminal law is insufficient to counter the risks to personal information protection brought about by this stage.The demand of generative AI for vast amounts of information inherently conflicts with the control and protection concepts of criminal law.Collecting personal information without the consent of the rights holder not only affects the exercise of citizens’right to self-determination of personal information but also brings the risk of privacy leakage,challenging the security and management order of personal information.The insufficiency of regulations under the traditional personal information protection model is becoming increasingly severe,and the determination of“serious circumstances”in the judicial interpretation of the crime of infringing upon citizens'personal information has fallen into a predicament.Therefore,efforts should be made from three aspects to enhance the criminal law's ability to respond to personal information protection in the new situation.Firstly,personal information disclosed publicly should be classified into three categories for protection:active disclosure,passive disclosure,and illegal disclosure.Secondly,the collection of personal information in criminal law should follow the principles of legality,propriety,and necessity.Finally,the type of personal information should be the first criterion for determining“serious circumstances”and the classification of personal information should be improved.
作者
蔡军
可欣雨
CAI Jun;KE Xinyu(Institute of Crime Control and Criminal Policy,Henan University,Kaifeng 475001,Henan,China)
出处
《昆明理工大学学报(社会科学版)》
2025年第6期1-12,共12页
Journal of Kunming University of Science and Technology(Social Sciences)
关键词
生成式人工智能
语料训练
个人信息
个人隐私
侵犯公民个人信息罪
generative AI
corpus training
personal information
personal privacy
crime of infringing upon citizens'personal information
