Digital governance of generative AI in higher education: authentic assessment as a pathway to student engagement and educational equity

Universities worldwide are struggling to govern students’ use of generative artificial intelligence (GenAI), yet it remains unclear whether institutional governance actually shapes meaningful learning. Existing student-level evidence is largely descriptive, and the little inferential work available suggests that mere awareness of AI policies has no effect on student conduct—leaving the effects of governance quality untested. Drawing on procedural justice, self-determination, and digital divide perspectives, this study develops and tests a model in which perceived GenAI governance quality shapes students’ perceptions of authentic assessment, which in turn relate to student engagement and perceived educational equity. Two-wave survey data were collected from 644 undergraduate and postgraduate students across research-intensive, teaching-oriented, and vocational-applied universities in eastern, central, and western China, and were supplemented by objective coding of institutional AI policies. Estimated with institution-clustered standard errors, governance quality was positively associated with perceived authentic assessment well beyond mere policy awareness, more than doubling the explained variance. Authentic assessment fully mediated the relationship between governance quality and student engagement and partially mediated its relationship with perceived educational equity, while digital access remained an independent predictor of equity. Findings indicate that AI governance becomes educationally meaningful primarily when students experience it as authentic assessment, whereas equity additionally requires attention to procedural and resource conditions. Implications for institutional AI policy design are discussed.