The psychological dynamics of AI in education: how bright traits and dark apprehensions relate to perceived academic performance and continuance

Generative artificial intelligence (AI) is increasingly embedded in higher education, yet the psychological factors associated with students’ continued use remain unclear. Integrating the Expectation-Confirmation Model, Two-Factor Theory, and Cognitive Appraisal Theory, this study examines how perseverance of effort, consistency of interest, technophobia, and creepiness relate to cognitive evaluations, perceived AI-assisted academic performance, and continuance intention. Survey data from 655 preservice teachers at four Chinese universities were analyzed using partial least squares structural equation modeling (PLS-SEM). The model explained 56.8% of the variance in continuance intention. Consistency of interest was positively associated with perceived usefulness, whereas perseverance of effort was associated with confirmation but not directly with perceived usefulness. Technophobia was positively associated with creepiness. Notably, creepiness showed a small positive association with perceived AI-assisted academic performance, measured through students’ self-reported confidence, assignment completion, and task efficiency, while its direct association with continuance intention was not statistically significant. These findings suggest that educational strategies for generative AI should consider both students’ psychological dispositions and perceived learning benefits, rather than assuming that technological discomfort necessarily discourages continued use. However, the cross-sectional, self-reported design does not establish causal effects or objectively improved academic achievement.