Bing chat for kidney stone management questions based on the AUA guidelines: a comparison of chatbot conversation...

Artificial intelligence (AI) technology will inevitably permeate healthcare. Bing Chat is an AI chatbot with different conservation styles. We evaluated each of these response mode answers regarding management of nephrolithiasis.

A total of 20 questions were created based on the AUA Surgical Management of Stones guidelines. Bing Chat's responses were evaluated across Precise, Balanced, and Creative conversation style chat modes by three physicians using the Brief DISCERN tool. Consensus scoring was employed to assess appropriateness, guideline adherence, empathy, recommendation for physician consultation, and inability to answer the inquiry. Responses were also assessed for their directness and the presence of superfluous information. Chat modes were compared using descriptive statistics as well as ANOVA, Chi-Squared tests, and Fisher exact tests.

The median Brief DISCERN Score in Precise, Balanced, and Creative modes were: 22, 21, and 21, respectively. There was no significant difference in Brief DISCERN scores between the three chat modes (p = 0.68). Guideline adherence by chatbot conversation style was similar (p = 0.37), as was response appropriateness (p = 0.62), directly answering the question asked (p = 0.26) and providing a recommendation to consult with a healthcare provider (p = 0.07). Creative and balanced modes outperformed precise mode when evaluating response empathy. Creative mode was more likely to include superfluous information and less likely to answer the question.

In its current iteration, Bing Chat provides low quality urologic healthcare information for nephrolithiasis queries, regardless of the conversation style utilized.