Abstract: [Purpose/Significance] The application of AI technology has become a trend of service innovation in public libraries. Exploring the AI's practical applications in serving vulnerable groups in public libraries and the ethical risk regulation path of these applications can help promote the formation of a responsible AI application ecosystem. [Design/Methodology] Through combining literature research and case analysis methods, this study elucidates the significance of AI application in serving vulnerable groups in public libraries, summarizes the main practice types of current AI application, analyzes the key ethical risks, and proposes regulatory path based on socio-technical system theory. [Findings/Conclusion] The main practice types include: resource localization and audio guidance services, intelligent robots reading assistance and educational services, and barrier-free smart services. There are ethical risks such as algorithm bias, technological dependence, trust crisis, and AI divide. The ethical risk regulation approaches include: strengthening librarians' awareness and practice of AI ethics; creating a friendly environment for vulnerable groups to learn and use AI technology; establishing AI ethical standards for the public library industry and creating platforms for AI ethical learning and practice communication; integrating AI accessible technology to provide trustworthy and inclusive AI services. [Originality/Value] This study introduces socio-technical systems theory, which can provide a framework and targeted strategies for identifying and regulating ethical risks in AI applications in public libraries.

Keywords: Public library, Vulnerable groups, Artificial intelligence ethic, Ethical risk, Artificial intelligence literacy, Socio-Technical system