Behavioral intention and actual use of ChatGPT in higher education

Abstract

Generative artificial intelligence (AI) tools such as ChatGPT are increasingly integrated into higher education, yet research emphasizes students’ intentions to adopt these tools rather than their actual use. This study examined whether behavioral intention predicts self-reported usage frequency. Using secondary survey data from higher-education participants (N = 751), a three-item behavioral intention composite and a 0–7 usage-frequency index were analyzed using descriptive statistics, Pearson correlation, and simple linear regression. Mean intention was M = 3.35 (SD = 0.91) and mean frequency was M = 3.22 (SD = 2.10). Behavioral intention correlated positively with usage frequency, r = .49, p < .001, and significantly predicted use, B = 1.12, SE = 0.07, β = .49, R² = .239, F (1, 748) = 235.04, p < .001. Findings support TAM and UTAUT while suggesting that institutional supports are needed to translate intention into sustained and responsible AI use in higher education.