Exploratory study compares GPT-4 and Gemini's performance in dental education assessments, highlighting accuracy and consistency.
Key Points
ChatGPT achieved an accuracy of 90% in MCQs and 85% in T/F questions, signaling its strong performance.
Gemini's accuracy varied between 60% and 70%, with the highest in SAQs at 70%, indicating room for improvement.
The Kappa coefficient showed ChatGPT's consistency at 0.754, while Gemini's was lower at 0.634, suggesting variances in reliability.
Findings emphasize the potential of AI in enhancing dental education, but highlight the need for thoughtful integration to maintain academic integrity.