Artificial intelligence (AI) chatbots may have the strong potential to transform how patients understand dental prostheses by mimicking human conversation, according to a study published on April 15 in BMC Oral Health.
Furthermore, these chatbots may significantly improve and replace traditional approaches to answering patients' frequently asked questions (FAQs), the authors wrote.
“AI chatbots show great potential in answering patient inquiries about dental prostheses,” wrote the authors, led by Hossein Esmailpour of the Guilan University of Medical Sciences School of Dentistry in Iran.
For this study, 31 FAQs related to removable and fixed prosthodontics were gathered from accredited organization websites and Google’s People Also Ask feature. Three chatbots -- ChatGPT 3.5, Microsoft Copilot, and Google Gemini -- were assessed. Each chatbot was asked the same set of questions on the same day without any prior prompts, they wrote.
Two prosthodontists then evaluated the responses using a modified Global Quality Score (GQS), assessing them for correctness, accuracy, and completeness to determine overall quality. Readability was assessed using the Flesch-Kincaid Grade Level (FKGL) and Flesch Reading Ease (FRE).
Among the chatbots, Google Gemini achieved the highest quality score (4.58 ± 0.5), significantly outperforming Microsoft Copilot (3.87 ± 0.89) (p = .004). In terms of readability, ChatGPT generated significantly more complex responses (10.45 ± 1.26), they wrote.
Furthermore, with the FRE’s assessment of how easy the responses were to understand, ChatGPT had the lowest score. These differences were statistically significant when compared to both Gemini (p < .001) and Copilot (p = .002). Gemini scored the highest in readability, with a significant advantage over Copilot.
However, further research is needed to explore how chatbots perform in realistic scenarios, including multiturn conversations and context-aware interactions, the authors added.
“These chatbots have high potential for replacement or improvement of traditional methods of responding to patient questions,” they concluded.
