THE ETHICAL GOVERNANCE OF GENERATIVE ARTIFICIAL INTELLIGENCE IN CHINESE EDUCATIONAL INSTITUTIONS

Abstract

Generative artificial intelligence (GenAI) is moving from occasional experimentation to routine use in teaching, assessment, research, student services, and educational administration. Its educational value is substantial, but so are the institutional risks created by inaccurate output, opaque decision processes, personal-data exposure, uneven access, uncertain authorship, and excessive dependence on commercial platforms. This conceptual policy analysis examines how Chinese educational institutions can govern GenAI in ways that are legally compliant, educationally defensible, and operationally workable. Drawing on human-centred AI ethics, sociotechnical and risk-based governance, Chinese data and GenAI regulation, national education policy, and recent education research, the paper argues that compliance alone does not provide an adequate institutional operating model. It develops the GUARDIAN Framework, organized around legitimate goals, user rights and human agency, assigned accountability, risk-tiered authorization, data protection, integrity and disclosure, inclusive and age-appropriate access, and continuous monitoring. The framework is intended for schools, vocational institutions, and universities, with safeguards calibrated to learner age, data sensitivity, decision consequence, and institutional capacity. The analysis concludes that responsible adoption depends less on the number of AI tools deployed than on the institution's ability to preserve professional judgment, credible learning, fairness, privacy, and avenues for human review throughout the technology lifecycle.