Awareness, attitudes, and educational use of artificial intelligence among medical students: a large cross-sectional survey

Artificial intelligence (AI) is increasingly influencing medical education and clinical practice. Understanding medical students’ awareness, attitudes, and usage patterns of AI is essential for guiding curriculum development and ensuring responsible integration into undergraduate medical training. A nationwide cross-sectional survey was conducted between November and December 2025 among undergraduate medical students in Türkiye. A structured questionnaire assessed demographic characteristics, AI awareness and frequency of use, educational and clinical applications, attitudes toward AI, ethical concerns, and future educational expectations. A composite AI Attitude Score was calculated using three Likert-scale items, and internal consistency was evaluated using Cronbach’s alpha. Descriptive and inferential statistical analyses were performed. A total of 1,346 medical students were included. Overall, 81.1% reported awareness of AI applications in medicine. AI tools were predominantly used for educational purposes (73.3%), whereas clinical or simulation-based use was limited (11.6%). The mean composite AI Attitude Score was 3.41 ± 0.61, indicating moderately positive perceptions (Cronbach’s alpha = 0.71). Ethical concerns were reported by 67.4% of participants. A majority (79.1%) expressed interest in receiving formal AI-related education. No significant differences were observed between preclinical and clinical students in overall attitudes or educational demand; however, clinical students reported significantly greater use of AI in clinical contexts (p < 0.001). Medical students demonstrate high awareness and generally positive attitudes toward AI; however, clinical integration remains limited and ethical concerns are prevalent. These findings underscore the need for structured, ethically grounded AI education within undergraduate medical curricula. Findings should be interpreted with consideration of potential self-selection bias due to voluntary online participitation.