Factors Influencing the Utilization of Large Language Models among Medical Students in Bangladesh: A KAP Study

Background Large Language Models (LLMs) have demonstrated remarkable potential in enhancing medical education. This study explored the knowledge, attitudes, and practices of undergraduate medical students in Bangladesh regarding LLM utilization. Methods This cross-sectional study, conducted from March to June 2024, assessed the knowledge, attitudes, and practices (KAP) of undergraduate medical students (MBBS) in Bangladesh. A convenience sampling method was used, with 1000 participants. A structured questionnaire, validated with acceptable Cronbach's alpha (knowledge: 0.703, attitude: 0.707, practice: 0.809), was distributed online via Google Forms. Data were analyzed using descriptive statistics, percentile-based thresholds, and multivariate ordinal logistic regression in R (version 4.3), with statistical significance set at P  < 0.05. Results Most participants exhibited poor knowledge (43%), negative or uncertain attitudes (78%), and low engagement with LLMs (37%). Male students and those from private institutions showed significantly higher knowledge, more positive attitudes, and greater utilization of LLMs than their counterparts. Ordinal logistic regression confirmed these associations, highlighting gender and institutional type as the key determinants. Conclusions These findings underscore the need for targeted interventions to improve AI literacy, address disparities in access, and effectively integrate LLMs into medical curricula. Faculty training, institutional support, and careful planning are essential for harnessing the benefits of LLMs while mitigating concerns about accuracy and over-reliance. Future research should explore the longitudinal trends and evaluate the impact of AI-based interventions on learning outcomes. This study provides valuable insights into the current state of LLM adoption in medical education in Bangladesh, and emphasizes the importance of equitable access, training, and integration to maximize the potential of these transformative technologies.