Cross-sectional associations of generative AI acceptance with self-efficacy and well-being among students recruited from seven Chinese universities

Abstract

Generative artificial intelligence (AI) is increasingly used in higher education, but its associations with students’ broader psychological outcomes remain uncertain. This cross-sectional study examined generative AI acceptance, general self-efficacy, and well-being among 495 students recruited from seven universities in China. Participants completed the Generative Artificial Intelligence Acceptance Scale, the New General Self-Efficacy Scale, and the WHO-5 Well-Being Index. Descriptive statistics, correlations, confirmatory factor analysis, and observed-variable path analysis were conducted in R. In unadjusted analyses, generative AI acceptance was positively associated with self-efficacy (β = 0.494, p < 0.001), and self-efficacy was positively associated with well-being (β = 0.217, p < 0.001). The direct association between AI acceptance and well-being was not statistically significant (β = −0.060, p = 0.240). The model accounted for 24.4% of the variance in self-efficacy and 3.8% in well-being. These coefficients were unadjusted for measured demographic characteristics. These findings should be interpreted within the context of the recruited non-probability sample. Universities should pair AI-related learning support with digital literacy, responsible-use guidance, and established student-support services.