Artificial intelligence self-efficacy among iranian medical sciences students: academic, behavioral, and socioeconomic correlates

Abstract Background Artificial intelligence (AI) is increasingly integrated into healthcare and medical education. This study assessed AI self-efficacy and its associated factors among medical sciences students with prior self-reported AI use. Methods In this cross-sectional study, 550 students from Tabriz University of Medical Sciences, Iran, were recruited using convenience sampling through an online questionnaire distributed via university social media channels and student networks. The questionnaire covered demographic, academic, socioeconomic, and digital media use characteristics, as well as the Persian version of the Artificial Intelligence Self-Efficacy Scale (AISES). Prior self-reported AI use was assessed using a Likert-scale item; students selecting “not at all” were excluded. Univariable and multivariable linear regression analyses were used to examine factors associated with AI self-efficacy. Results The mean AISES score was 69.80 ± 13.34 on a standardized 0–100 scale. Educational level was associated with AI self-efficacy, with bachelor’s (β = 15.34, 95% CI: 8.05–22.64), master’s/PhD (β = 12.63, 95% CI: 6.88–18.48), and professional-doctorate students (β = 18.63, 95% CI: 13.71–23.55) having higher scores than residents. Higher GPA (β = 3.62, 95% CI: 2.44–4.80) and employment (β = 5.25, 95% CI: 1.03–9.49) were also associated with higher AI self-efficacy. Compared with < 1 h/day, educational digital-media use of ≥ 3 h/day was associated with higher AI self-efficacy, whereas recreational use of ≥ 3 h/day was associated with lower scores. Age and gender were not independently associated with AI self-efficacy. Conclusions AI self-efficacy was associated with academic, socioeconomic, behavioral, and educational characteristics among medical sciences students. These cross-sectional associations do not establish causality; longitudinal studies are needed to clarify temporal direction.