生成式AI赋能图书馆中的读者信息分类保护

Generative AI-enabled Classified Protection for Library Reader Information

生成式AI的技术发展与应用推广为图书馆数字化建设提供新契机,但也给读者信息安全带来挑战,表现为不当处理读者信息的规模更庞大、行为更复杂且场景更多样。而现有的读者信息保护制度对此存在应对局限,生成式AI所涉读者信息承载利益呈现失衡状态。为解决这一问题,需要结合《公共图书馆法》中的读者信息类型形成分类保护规则,以此平衡生成式AI对读者信息的利用诉求与保护需要。其中,读者隐私信息需要强化保护,生成式AI应当严格落实保护隐私信息的知情同意与目的限制等规定;读者个人信息需要有限保护,生成式AI可以灵活遵循个人信息处理规则;读者借阅信息需要弱化保护,生成式AI对借阅信息的处理活动得到鼓励与支持,但需履行防止滥用的信息保护义务。

The development and application of generative AI provides new opportunities for library digitization,but it also brings challenges to the security of reader-related information,which manifests itself as the mishandling of reader information on a larger scale,with more complex behaviors and diverse scenarios.However,the existing reader information protection system has limitations,and there is an imbalance in the interests of reader information involved in generative AI.To solve these problems,it is necessary to develop classified protection rules according to the types of reader information defined in the Law on Public Libraries,so as to balance the claims of generative AI on patron information with the need to protect user information.Among them,stronger protection should be given to readers’private information,for which generative AI must strictly implement the rules of informed consent and purpose limitation;limited protection should be given to readers’personal information,the processing rules of which generative AI can flexibly follow;lesser protection should be given to library loan information,the processing of which by generative AI is encouraged,with obligations to prevent abuse of information protection.