Leading Teams in the Digital Age: Team Technology Adaptation in Human-Agent Teams

Larson, Lindsay Elizabeth

doi:https://doi.org/10.21985/n2-xyzt-az21

Work

Leading Teams in the Digital Age: Team Technology Adaptation in Human-Agent Teams

Public

Download PDF

Download All Files (.zip)

This dissertation imagines the near future of teamwork, when AI agents will join teams,interacting, collaborating, and completing tasks as a team member. Broadly, I seek to answer the questions: how do humans integrate a new AI teammate onto their team, and how does the AI teammate’s function influence this integration process? Further, given the integral role of team cognitive processes: how do human team members adjust their transactive memory systems (TMSs) to accommodate agent teammate newcomers? To better understand this phenomenon, I propose the concept of team technology adaptation and elaborate a developmental stage model to explain how teams adjust their mental representations and interactions in response to the introduction of an AI teammate. I studied the effects of an AI newcomer on team functioning in a series of two studies. Study 1 used a sample of 365 MTurk workers to validate measures of cognitive processes derived from my stage model. Study 2 was a laboratory experiment including 63 teams (149 individuals) who adjusted to the addition of an AI newcomer that was randomly assigned to support teamwork, taskwork, or both teamwork and taskwork. The AI, named “Vero”, was implemented using a Wizard of Oz methodology, with a confederate, pre-validated team function prompts, and visual animations. Teams performed three parallel problem solving and creative thinking tasks, first without an AI, and then in two subsequent rounds with Vero. Key findings are presented in Table 1. In this dissertation, I introduce the concept of team technology adaptation, contributing to our current understanding and future exploration of how humans experience agentic team technologies and how it affects team processes and states. Further, I demonstrate that agent teammates occupy a unique position within the human-agent TMS that is important in predicting team behavioral process. Results also provide evidence for the important role of AI teammate schema in the development of human-agent TMS. Finally, I find that it is not the presence of an agent teammate that only completes teamwork processes, but rather the lack of taskwork processes, that lead to less positive processes and outcomes in human-agent teams. These results have practical implications for how leaders prepare their teams to integrate AI teammates onto teams.

Creator