Presentation
Call 9-1-1 or Call ChatGPT?: Human-AI Teaming for Two Simulated Medical Emergencies: An Experimental Study
DescriptionThis study aims to compare human-human teaming (HHT) and human-AI teaming (HAT) in response to two simulated medical emergencies using manikins: choking and severe bleeding. Especially, our study uses ‘ChatGPT’, one of the most advanced large language models (LLMs), as an AI agent who provides instructions for the medical situations. Our study is focused on evaluating medical treatment performance (task completion time, expert rating), trust on another teammate (human vs. ChatGPT), self-efficacy and perceived workload. Regarding task completion time, there was no significant main effect of teammate type. However, there was significant main effect of task type. Also, there was a significant interaction effect between task type and teammate type. Regarding expert rating, HHT showed significantly higher rating than HAT. HHT was also given higher team trust than HAT. Regarding NASA TLX, a significant interaction effect was found between task and teammate with HHT for choking task showing the highest perceived workload. Implications of the current findings and future research directions are discussed.
Contributors
Alternate Presenter
Event Type
Lecture
TimeTuesday, October 14th4:50pm - 5:10pm CDT
LocationGrand B
Human AI Robot Teaming (AI)
Similar Presentations