Psychometric properties of the Persian version of the artificial intelligence self-efficacy scale (AISES) in medical sciences students

As artificial intelligence (AI) becomes increasingly integrated into healthcare, assessing medical students’ confidence in using AI tools is essential. This study aimed to translate, adapt, and psychometrically validate the Persian version of the Artificial Intelligence Self-Efficacy Scale (AISES) among Iranian medical sciences students, ensuring its suitability for assessing AI-specific self-efficacy in this population. A cross-sectional study was conducted in late 2023 using a convenience sample of 501 students recruited via an online survey. Content and face validity were evaluated both qualitatively and quantitatively. Construct validity was examined using Exploratory and Confirmatory Factor Analysis (EFA and CFA). Item Response Theory (IRT) with the Graded Response Model was applied to assess item-level discrimination. Convergent and discriminant validity were tested using the Average Variance Extracted (AVE) and the Fornell–Larcker criterion. Reliability was assessed with Cronbach's alpha, McDonald's omega, and Composite Reliability. EFA supported a four-factor structure consistent with the original AISES, explaining 77.27 % of the variance. CFA confirmed model fit (CFI = 0.944, RMSEA = 0.057). IRT results showed high item discrimination (α = 1.06–3.78) and logically ordered thresholds. The Test Information Function showed the highest precision in the lower-than-average range of AISE. All reliability coefficients exceeded 0.80. AVE values confirmed convergent validity, and discriminant validity was supported by the Fornell-Larcker criterion. The Persian AISES is a valid, reliable, and culturally adapted tool for assessing AI self-efficacy in Iranian students. Its ability to identify low-confidence learners supports targeted curriculum design, early intervention, and more equitable AI-based learning. • Practical Use in Education and Research: The Persian AISES can help educators and researchers assess students' readiness for AI, spot those who need extra support, and design targeted learning interventions. • Task-Specific Advantage: As a task-specific self-efficacy tool, AISES captures AI-related confidence more precisely than general self-efficacy scales, in line with Bandura's recommendations. • Cultural and Linguistic Adaptation: The scale was carefully translated and adapted for Iranian medical sciences students, ensuring clarity, cultural fit, and relevance to the local context. • Robust Psychometric Testing: Evaluation combined classical methods with Item Response Theory, including face/content validity, EFA, first- and second-order CFA, and convergent/discriminant validity checks. • Four Key Dimensions Identified: EFA confirmed four factors—Assistance, Anthropomorphic Interaction, Comfort with AI, and Technological Skills—explaining 77.27 % of variance, consistent with the original AISES structure.