The Challenge of Generative Artificial Intelligence to Personal Information Protection and Its Risk Regulation
The technical characteristics of Generative AI pose a challenge to the personal information protection system established by China's Personal Information Protection Law,including:although Generative AI has solved the problem of pre-training language models with large-scale unlabeled data,the adopted technical route also makes Large Language Models(LLMs)a complete technical black box,making it difficult for developers to comply with the rules of informed consent for personal information processing;whether based on the principle of purpose limitation or the theory of scenarios,the technical characteristics of Generative AI make it difficult to meet the statutory requirement of processing personal information within a"reasonable scope"in accordance with the law;the technical characteristics of Generative AI make it possible for LLMs to infringe on the sensitive personal information rights and individual privacy rights of information subjects at both the input and output ends.We should base our basic risk regulatory philosophy of"inclusive prudence"on this and take measures such as adjusting the application of informed consent rules in the field of Generative AI,reshaping the rules for processing publicly available personal information in the field of Generative AI,and establishing administrative regulatory measures for protecting the personality rights of personal information in the field of Generative AI,to achieve a balance between innovative technology development and personal information protection.
万方数据
维普
