AI literacy and AI-assisted learning appraisal among undergraduate students in Vietnam: A cross-sectional higher-order structural equation modeling study

Students’ evaluations of AI-assisted learning may share a broader common appraisal. This study evaluates whether AI-assisted learning appraisal can be represented as a higher-order construct reflected by trust in AI, perceived usefulness, continued-use orientation, and perceived learning benefits, and examines its association with self-reported AI literacy. A total of 519 consenting survey responses were collected through classroom convenience sampling at a single Vietnamese university and analyzed using confirmatory factor analysis and structural equation modeling. The findings provide qualified support for the higher-order representation, although substantial overlap limits the empirical distinctiveness of its first-order dimensions. AI literacy was positively associated with appraisal across response-quality samples and estimation methods, although the standardized coefficient declined from 0.867 in the full sample to 0.723 after stricter response-quality screening. The higher-order construct offers a parsimonious summary of learning-related evaluations, but its distinctiveness from general AI favorability remains untested. The cross-sectional convenience design and unverified uniqueness of respondents across the two collection periods limit causal interpretation and generalizability, and leave response independence uncertain.