Presentation
Automating the Identification of Team-Based Leadership and Teamwork Characteristics Using LLMs
DescriptionThe primary aim of this research was to explore and validate the application of a large language model (LLM), specifically GPT-4, in automating the identification of team-based emergent leadership (Maese et al, 2023) and critical teamwork characteristics within the team communication of a large multi-team experiment. This study sought to harness the analytical context-driven capabilities of LLMs to discern and categorize key team dynamics and leadership attributes from a dataset derived from a large experimental team-based Minecraft simulation environment. The objectives were twofold: first, to automate the detection of team leaders; and second, to identify and analyze ten essential characteristics associated with team performance, including, but not limited to, motivating language, planning, coordination, transactive memory systems, and compensatory helping. Overall, the study aimed to offer novel insights into the automation of team analysis and contribute to the broader understanding of applying artificial intelligence in organizational and team dynamics research.
Contributors
Event Type
Lecture
TimeWednesday, September 11th1:30pm - 1:50pm MST
LocationFlagstaff
Human AI Robot Teaming (AI)