Abstract: [Purpose/Significance] This study aims to analyze the influence mechanism of various anthropomorphic cues and user perceptions on satisfaction with government chatbots, thereby offering scholarly references for advancing the application of artificial intelligence in government affairs and enhancing the capabilities of digital public services. [Design/Methodology] Using a mixed research design, we collected user data through experiments and interviews, and conducted a qualitative research analysis based on grounded theory. Quantitative research data were then obtained through eight types of human-AI interaction scenario questionnaires for Structural Equation Modeling（SEM）testing. [Findings/Conclusion] The qualitative study identifies nine main categories and twenty-seven subcategories influencing user interactions with government chatbots, along with the relational structure and storyline of the core categories. The quantitative study reveals that all four types of anthropomorphic cues in government chatbots significantly affect social presence. However, only linguistic and emotional anthropomorphization（systematic information processing cues）significantly influence users' perceived warmth. Both social presence and perceived warmth positively impact users' functional and emotional value perceptions, with perceived emotional value exerting a greater influence on user satisfaction than perceived functional value. [Originality/Value] This study examines the multidimensional factors affecting user satisfaction with government chatbots, and verifies the complex relationships among various anthropomorphic cues, user perceptions, and satisfaction in human-AI interactions. The findings provide both theoretical and practical insights for advancing the application of artificial intelligence in government affairs.

Keywords: Government chatbots, Anthropomorphism, Perceived warmth, Social presence, User satisfaction