Correlates of University Students' AI Usage Intention for Learning: A Secondary Analysis of Indonesian Students

Abstract

Artificial intelligence (AI) is now widely used in higher education, but adoption indicators do not by themselves show which factors are uniquely associated with students’ intention to use AI for academic work. This study reanalysed publicly available aggregate results from a 2024 survey of 535 students at 152 Indonesian higher education institutions in 20 provinces. The source reported item means, AI-tool frequencies, and a correlation matrix for performance expectations, challenges, attitudes, perceptions, and a construct originally labelled motivational intention. For clarity, the present article reports this outcome as AI usage intention without changing its items, scores, or correlations. Equal-weighted construct means were reconstructed from the published item means, and standardized multiple regression was derived from the published correlation matrix. The four-predictor model accounted for approximately 55.4% of the variance in AI usage intention. Attitudes had the strongest unique association (β = .361), followed by perceptions (β = .277) and performance expectations (β = .242), all p < .001. Challenges correlated with usage intention at the bivariate level (r = .310) but made no unique contribution after overlap with the other constructs was considered (β = .012, p = .720). ChatGPT was the most frequently reported primary tool (47.5%). The findings represent statistical associations rather than causal effects. Universities should combine access with discipline-specific AI literacy, transparent assessment rules, output verification, and learning activities that preserve student judgement.