征信

数字时代，生成式人工智能（AIGC）可收集并处理海量个人信息，但其如何履行个人信息删除义务并保护个人信息权益存在巨大争议。在绝对与相对删除义务之间，因绝对删除义务不符合比例原则且存在客观行使障碍，应转向合目的性和现实性的相对删除义务。相对删除义务由实体和程序标准所组成。其中，实体标准应在区分未公开和已公开个人信息的基础上，以风险控制为宗旨，分别界定AIGC的删除义务及与之相对应的删除主体权利和监督者权力；程序标准以事前、事中、事后作为类型化方式，分别释明三大主体删除程序中的相关权责。实体与程序标准相结合，使得AIGC的相对删除义务得以形成动态、高效的信息管理机制，为个人信息保护提供坚实的法律和技术保障。

In the digital age,generative artificial intelligence(AIGC)can collect and process a massive amount of personal information. However,there is a significant controversy surrounding how it fulfills the obligation to delete personal information and protects the rights and interests of personal information. Between absolute and relative deletion obligations,due to the impracticality of the absolute deletion obligation for not meeting the proportionality principle and the existence of objective exercise barriers,there should be a shift towards a purposeful and realistic relative deletion obligation. The relative deletion obligation consists of substantive and procedural standards. The substantive standards should,based on the distinction between non-disclosed and disclosed personal information,aim at risk control,and respectively define the deletion obligations of AIGC and the corresponding rights of deletion subjects and the powers of supervisors. The procedural standards,categorized by pre-event,during-event,and post-event,clarify the relevant rights and responsibilities in the deletion procedures of the three major subjects. The combination of substantive and procedural standards allows for the formation of a dynamic and efficient information management mechanism for the relative deletion obligation of AIGC,providing solid legal and technical guarantees for personal information protection.