Chatbots and Medical Sciences Education: A Review of Applications and Challenges

Background and Aims: Recent advancements in artificial intelligence, particularly large language models (LLMs) such as ChatGPT, have significantly influenced medical education. These technologies are capable of supporting various educational domains, including fundamental medical concept learning, differential diagnosis, clinical scenario simulations, and problem-based learning (PBL). Given the ever-expanding nature of medical knowledge, LLMs, with their ability to process vast amounts of information, have the potential to enhance students' comprehension of complex topics. However, the implementation of chatbots in medical sciences education presents both opportunities and challenges that warrant careful examination. Methods: This study employed a narrative review approach by systematically analyzing relevant literature from PubMed, Google Scholar, and MBC. Articles published between ۲۰۲۰ and ۲۰۲۵ were selected based on their direct relevance to the topic, methodological rigor, and citation impact. Following an initial screening, ۲۵ studies were identified for in-depth analysis and synthesis. Results: Findings indicate that chatbots contribute to medical education in four primary areas: designing educational content and clinical scenarios, simulating physician-patient interactions, assessing student performance and providing feedback, and facilitating self-directed learning. Despite these benefits, several challenges persist, including the relative accuracy of chatbot-generated medical information, the inability to perform complex clinical reasoning, limitations in evaluating practical skills, and concerns regarding data privacy in patient-related applications. Additionally, ethical considerations related to excessive reliance on artificial intelligence for medical sciences training have been identified. Conclusion: While chatbots play a valuable role in medical education, they should not be considered a substitute for traditional teaching methods. Instead, they serve as complementary tools that enhance learning, foster student engagement, and enable realistic clinical scenario simulations. To mitigate their limitations, measures such as faculty oversight of chatbot-generated content, integration with authoritative