No Time for ChitChat? Artificial intelligence Chatbots as a tool to identify research priorities in hip and knee arthroplasty.

Abstract

Abstract Background: Artificial intelligence (AI) Chatbots, such as ChatGPT3, have gained attention in medical and non-medical domains. Their ability to identify research gaps in orthopaedics is yet to be tested. Aims: This study aimed to assess the application of three AI Chatbots to identify research questions in hip and knee arthroplasty in comparison to an existing research prioritisation consensus method. Methods: Three Chatbots, ChatGPT3, Bing and Bard were prompted to identify research questions in hip and knee arthroplasty. Two authors independently compared the responses to the 21 research priorities for hip and knee arthroplasty established by the James Lind Alliance (JLA). Any discrepancies were discussed with senior authors. Results: ChatGPT3 successfully identified to 15 (71%) priorities. Bard, nine (42%) priorities, while Bing identified eight (38%). The Chatbots identified further questions that were not stated in the JLA exercise (ChatGPT3: 12 questions; Bard: 14 questions; Bing: 11 questions). All three Chatbots failed to identify five (24%) of the JLA research priorities. Conclusions: This study reports the first evidence of the potential adoption of AI Chatbots to identify research questions in hip and knee arthroplasty. This may potentially represent a valuable adjunct in improving efficiency of research prioritisation exercises.