Evolving resource use for self-directed learning in physiology among first-year medical students in a classroom setting.

Self-directed learning (SDL) is integral to medical education. It helps in fostering critical thinking, independence, and problem-solving skills. With advancements in technology, digital tools like search engines, interactive content, and large language model (LLM) chatbots have become supplementary tools to traditional materials such as textbooks. However, limited data exist on SDL resource preferences among medical students in India since the inception of LLMs like ChatGPT. To address this, 64 medical students participated in a classroom-based SDL session on anemia. Students freely selected resources during a 40-min preparation period, followed by 20 min of writing responses to a total of five higher-order knowledge questions (i.e., questions started with "Explain why"). Postsession, they anonymously reported their resource use. Among 63 valid responses, 46.03% used one resource, 39.68% used two, and 14.29% used three. Search engines (61.9%) and LLM chatbots (60.32%) were the most frequently used, followed by textbooks (26.98%), with less reliance on notes, journals, and videos. Hence, there is a growing preference for search engines and LLM chatbots as an educational tool in self-directed learning in a classroom setting.NEW & NOTEWORTHY This study reports a shift in educational resource use for self-directed learning (SDL) among medical students after introduction of large language model (LLM) chatbots. Students use multiple sources, with digital tools like search engines and LLM chatbots nearly matching each other as the most preferred resources. Despite the accessibility and efficiency of digital tools, traditional resources like textbooks remain relevant, though less frequently chosen.