Potential Use of Artificial Intelligence (Chatbots) for Assessing Patient Satisfaction in Endodontics- Cross-sectional Comparative Study

Abstract

Aim: To evaluate and compare patient satisfaction with responses generated by Generative Pre-trained Transformer 4 (ChatGPT-4) and Generative Multimodal Intelligence 2 (GEMINI-2) to frequently asked questions in the Outpatient Department of Conservative Dentistry and Endodontics. Methodology: Seventy literate patients were enrolled over seven consecutive days. Each patient posed one question about their symptoms or treatment plan to both AI models using their personal device. After receiving AI responses, the treating clinician provided the correct explanation. Patients scored their satisfaction with each AI response using a five-point Likert scale (1 = Not at all satisfied; 5 = Very much satisfied). The Kolmogorov–Smirnov test confirmed non-normal data distribution, and the Mann–Whitney U test was used for group comparison. Results: The mean satisfaction score for ChatGPT-4 was 3.91 and for GEMINI-2 was 4.30. The Mann–Whitney U test revealed a statistically significant difference (U = 1829.500, P = .005), with GEMINI-2 receiving significantly higher patient satisfaction scores. Conclusion: GEMINI-2 demonstrated better patient satisfaction than ChatGPT-4 in an endodontic outpatient setting. AI chatbots show promise in supporting patient communication but cannot replace clinical expertise.