Assessing ChatGPT's Performance Compared to Dental Students in an Undergraduate Pediatric Dentistry Exam: Comparative Cross-Sectional Study

Objective: This study aimed to assess ChatGPT's knowledge in the field of pediatric dentistry by comparing its performance with that of 5th-year dental students. Material and Methods: The theoretical and clinical knowledge of 5th-year dental students 52 enrolled in the Turkish program and 49 in the English program was evaluated using a 310-item multiple-choice question exam. The same questions were presented to ChatGPT in both languages. A representative student with an average performance was selected from each program to statistically compare success rates with ChatGPT. Pearson chi-square and McNemar tests were used for statistical analysis. Results: No significant difference was observed in the correct answer rates between languages for either ChatGPT (p=0.497) or the representative students (p=0.327). While there was no significant difference between the Turkish program's representative student and ChatGPT (p=0.282), ChatGPT outperformed the representative student in the English program (p=0.004). The overall correct answer rates of students in the Turkish and English programs were 60.44% and 56.48%, respectively, with no statistically significant difference (p=0.686). Conclusion: ChatGPT demonstrated a performance comparable to that of senior dental students in pediatric dentistry exams. These findings suggest that Large Language Models may hold potential as educational support tools in dental education.