چکیده: (133 مشاهده)
Large Language Models (LLMs) are transforming dental education and practice by supporting clinical decision-making, administrative automation, and academic assessments. This review synthesizes 12 studies (May 2024–June 2025) evaluating LLMs, including ChatGPT, Gemini, and Claude, on dental board and academic examinations using a modified Population, Intervention, Comparison, Outcome (PICO) framework to assess accuracy, reliability, comprehensiveness, and reasoning quality. A narrative review of the literature was conducted, identifying relevant articles from PubMed, Scopus, Google Scholar, and arXiv. Twelve studies were selected for inclusion based on their relevance to the topic. LLMs achieved acceptable accuracy on multiple-choice questions, often surpassing human benchmarks, though performance varied by model, question type, and language. They excel in factual recall and exam preparation, particularly in resource-limited settings, but struggle with clinical reasoning and text-based formats. LLMs show significant potential for enhancing dental education, especially in standardized assessments, but require standardized evaluation frameworks, diverse question formats, and ethical guidelines to address limitations in practical and visual applications for effective integration into dental curricula.
نوع مطالعه:
Review article |
موضوع مقاله:
عمومى دریافت: 1404/4/16 | پذیرش: 1404/5/30 | انتشار: 1404/6/16