Could Freely Available Artificial Intelligence Chatbots Simulate a Multidisciplinary Workshop to More Efficiently Cogenerate Ideas?

Abstract

OBJECTIVE: We aimed to simulate a multidisciplinary workshop using artificial intelligence (AI) chatbots, to broaden and accelerate the cogeneration of ideas for a complex clinical research question. DESIGN: Proof-of-concept observational study. METHODS: Three freely available chatbots (ChatGPT 3.5 August 3 Version, Bing Creative Mode, and Google Bard v2023.07.13) were prompted on September 18, 2023, with an identical series of prompts to simulate a multidisciplinary workshop. Different knowledge users across clinical, research, policy, organizational, and leadership domains were suggested by the chatbots and then included. An appreciative inquiry approach integrated clinician and researcher expertise to provide solutions. RESULTS: All 3 chatbots suggested an online platform or curriculum for pain science education in schools. Each chatbot offered unique responses that were beneficial to reflect upon during idea generation. Key limitations of this approach, such as output bias and generalizability, and future research directions for using AI chatbots to simulate workshops were explored. CONCLUSION: Using AI chatbots to simulate a multidisciplinary workshop may be a useful way to efficiently incorporate a wide range of voices during idea generation. JOSPT Open 2024;2(3):225-232. Epub 17 April 2024. doi:10.2519/josptopen.2024.1133

Ähnliche Arbeiten