When Learning Reinforces Inertia: Organizational Conditions of AI Adoption in the Public Sector

This study examines how organizational information environments shape AI adoption intention in the public sector. Based on survey data from 1068 civil servants across 31 provincial-level jurisdictions in China, it develops and tests a framework incorporating organizational fit, institutional inertia, perceived workload, and organizational learning support. The results show that organizational fit facilitates AI adoption, whereas institutional inertia and perceived workload significantly inhibit it. Organizational learning support exhibits context-dependent moderating effects: it strengthens the positive role of fit and mitigates workload-related resistance, but paradoxically amplifies the negative impact of institutional inertia in highly rigid organizations. Heterogeneity analyses further reveal systematic variations across regions, job types, administrative levels, and position ranks. By conceptualizing AI as an embedded organizational information-processing system, this study contributes to understanding AI adoption dynamics in government organizations and highlights the need for context-sensitive learning strategies.