Student use of generative AI and perceived institutional support across three linguistic contexts

Student use of generative artificial intelligence (AI) raises questions about how everyday academic practices relate to perceived institutional support. This exploratory cross-sectional survey used Indonesian, English, and Hungarian questionnaire versions sharing a common set of core items. The harmonised dataset included 186 respondents, mainly residents of Indonesia and Hungary. Overall, 78.0% reported using AI several times per week or daily (95% CI [71.5, 83.3]), 96.8% reported using ChatGPT (95% CI [93.1, 98.5]), and 79.6% reported perceived learning improvement (95% CI [73.2, 84.7]). Institution-funded access and clearer policies were the most frequently requested forms of support. In the full sample, the Indonesia–Hungary contrast was larger for professor encouragement (|r rb | = 0.549) than for perceived learning improvement (|r rb | = 0.050). However, governance-related contrasts attenuated in the small English-language subsample. Country, language, recruitment route, and response timing overlapped, and measurement invariance was not established. These findings describe sampled students’ perceptions and do not identify independent country effects, institutional policy quality, or a temporal lag between practice and governance. We use practice-governance decoupling as an interpretive description of the coexistence of routine reported use and uneven perceived support. A four-component heuristic distinguishes survey findings from literature-informed interpretation and proposed outcomes for future evaluation.