Exploring artificial intelligence usage preferences among medical students in Iran: a cross-sectional study

Abstract Artificial intelligence (AI) is increasingly integrated into medical education, offering opportunities to enhance learning, research, and clinical preparation. Although previous studies have explored medical students’ attitudes and acceptance of AI, evidence regarding their AI usage preferences remains limited, particularly in developing countries where AI integration into medical education is still evolving. Understanding these preferences is essential for designing context-appropriate educational strategies and responsible AI integration. This study aimed to examine the preferences and patterns of AI usage among students at Shiraz University of Medical Sciences (SUMS), Iran. This descriptive cross-sectional study was conducted in 2025 among undergraduate and postgraduate students at SUMS. Using stratified random sampling, 384 students were recruited. Data were collected between July and October 2025 using a validated self-administered questionnaire (AI-UPSS; Cronbach’s α = 0.935), assessing AI usage across eight domains. Responses were rated on a 4-point Likert scale. Data were analyzed using one-sample t-tests, independent t-tests, and one-way ANOVA in SPSS version 24, with a cutoff mean score of 2.5. The highest mean preference scores were observed in Research & Writing (Mean = 2.66 ± 0.79, p < 0.001) and Self-learning/Education (Mean = 2.61 ± 0.81, p = 0.010). Moderate preferences were found for Curiosity & Entertainment (Mean = 2.36 ± 0.81, p = 0.001) and Health Counseling (Mean = 2.40 ± 0.96, p = 0.044). The lowest scores were reported in Design & Programming (Mean = 1.54 ± 0.79, p < 0.001) and Art (Mean = 1.68 ± 0.77, p < 0.001). Significant gender differences were observed in Art ( p = 0.010), Self-learning ( p = 0.027), and Health Counseling ( p < 0.001). GPA was significantly associated with Content Development ( p = 0.004), Design & Programming ( p = 0.031), and Art ( p = 0.032), with higher usage among students with lower GPAs. Undergraduate students showed significantly higher preferences across several domains, including curiosity and health counseling ( p < 0.01). Medical students primarily use AI for academic, research, and self-directed learning purposes, while engagement with technical and creative applications remains limited. AI usage preferences are influenced by academic level, gender, and GPA rather than age or field of study. These findings underscore the need for structured, level-specific AI education to promote effective and ethical AI integration in medical curricula.