Evaluation of ChatGPT as a Multiple-Choice Question Generator in Dental Traumatology

Authors : Yağız Özbay

Pages : 235-238

Doi:10.37990/medr.1446396

View : 45 | Download : 72

Publication Date : 2024-05-16

Article Type : Research Paper

Abstract :Aim: This study aims to evaluate the ability of ChatGPT-4 to generate clinical case-based multiple-choice questions in dental traumatology. Material and Method: International Association of Dental Traumatology guidelines for the management of traumatic dental injuries were introduced to ChatGPT-4 as an information source and ChatGPT-4 was prompted to 20 generate questions in fractures and luxations, avulsion of permanent teeth, injuries in the primary dentition. Sixty questions in total were generated by ChatGPT and the questions were evaluated by 2 endodontists using a 3-point scale. The One-way analysis of variance and Post Hoc Tukey test were used to analyze the data and the significance was P<0.05. Results: The average time to generate 20 questions was 1 min 55 sec. It was noted that 52% of the questions were usable without modification or with minor changes, while 28% were incorrect or completely useless. Conclusion: Despite the flaws, ChatGPT can be useful for creating multiple-choice questions in dental traumatology after a rigorous evaluation, elimination, and development procedure.
Keywords : artificial Intelligence, clinical practice guidelines, dental education, endodontics, treatment planning