Stakeholder Experiences of Artificial Intelligence in University Management at Two Chinese Universities in Hubei Province

Abstract

Research on artificial intelligence (AI) in higher education has concentrated on teaching and learning, leaving the administrative core of the university comparatively under-examined. This study asks how AI technologies are experienced, applied, and governed within the management systems of two public universities in China, and how AI-generated data enters managerial decision-making. A qualitative phenomenological design was applied. Six analytical constructs, including AI infrastructural embeddedness, administrative process integration, perceived administrative efficiency, human readiness, AI governance assurance, and decision augmentation, were derived from an integrated theoretical framework combining systems theory, diffusion of innovation and the technology–organisation–environment framework, human capital and AI capability theory, institutional theory, and socio-material and human-centred AI perspectives. Four instruments were used, semi-structured interviews with 15 participants (five administrators, four IT staff, three faculty members, three students), two focus group discussions (n = 10), analysis of eight institutional documents, and six non-participant observation sessions. Data were analyzed through open coding, category development, thematic synthesis, and interpretative phenomenological analysis, then triangulated across instruments. The findings showed that AI is experienced not as an innovation under trial but as normalized institutional infrastructure. Embedding is role-differentiated rather than uniform where administrators encounter AI through reporting and planning interfaces, IT staff through system architecture, faculty through learning analytics, and students through service platforms. Efficiency gains are consistently reported but bounded by interpretive rather than operational competence. Governance and privacy assurance operate as conditions of perceived value rather than as secondary compliance concerns. Across all participant groups, AI is positioned as advisory rather than autonomous. The two sites diverge consistently, University A integrates AI operationally, University B analytically, within a shared trajectory. The study specifies six theoretically grounded constructs, each with defined sub-dimensions, derivation, and operationalization, for research on AI in university management. It provides evidence of a later adoption stage than most of the literature describes and sets out a nomological network and measurement agenda for subsequent quantitative work.