Frequency and purpose of artificial intelligence tool use and their associations with academic motivation and self-efficacy among medical students: a cross-sectional study

Abstract Background Artificial intelligence (AI) tools are increasingly used by medical students for academic purposes, yet how patterns of such use relate to academic motivation and self-efficacy remain insufficiently characterized, particularly in the Iranian medical education context. This study examined the frequency and purpose of AI tool use among medical students and their associations with academic motivation and academic self-efficacy. Methods A descriptive-analytical cross-sectional study was conducted among medical students at Kashan University of Medical Sciences, Iran, in 2026. Using proportionate stratified sampling across four educational stages, 111 students completed a questionnaire package comprising a researcher-developed AI-use instrument (separate frequency and purpose scores), the Persian Academic Motivation Scale, and the Persian College Academic Self-Efficacy Scale. Associations between the two AI-use scores and five outcomes (self-efficacy, intrinsic motivation, extrinsic motivation, amotivation, and composite positive motivation) were examined using Spearman correlation with Bonferroni correction, followed by five multiple linear regression models adjusting for age, sex, educational stage, and GPA. Results Conversational AI tools, most notably ChatGPT, were used by nearly all participants. Both AI-use frequency and purpose were positively correlated with academic self-efficacy, intrinsic motivation, extrinsic motivation, and positive motivation, and negatively correlated with amotivation (Spearman ρ = 0.46–0.77; all p < 0.001, surviving Bonferroni correction). In adjusted models, AI-use purpose remained independently associated with all five outcomes (standardized β = 0.32–0.69), whereas AI-use frequency remained independently associated only with self-efficacy (β = 0.55, p < 0.001). GPA was independently associated with intrinsic motivation, extrinsic motivation, and positive motivation, but not self-efficacy or amotivation. No significant differences by sex or educational stage were observed. The five models explained 40% to 61% of outcome variance. Conclusions The purpose for which medical students use AI tools showed more consistent independent associations with academic motivation and self-efficacy than frequency of use alone, suggesting that how AI is used may be more relevant to these outcomes than how often it is used. As an observational, cross-sectional study, these findings describe association rather than causal sequence, and contribute context-specific evidence on AI use and its psychological correlates among medical students in Iran.