Hybrid Teaming with Generative AI: Role Dynamics and Their Consequences for Task Performance

Abstract

This study investigates how the integration of a task-specific responsive GenAI agent influences team role distribution and task execution. Building on collaboration and role theory, we conducted a qualitative experiment with six student teams, comparing experimental groups with AI access and control groups working with search engines. The results show that the effectiveness of GenAI depends on how critically and collaboratively teams engage with it. Teams that over-relied on AI - either due to uncertainty or lack of initiative - produced weaker outcomes than control groups, whereas those who challenged and contextualized AI outputs achieved higher performance. Furthermore, the presence of GenAI altered team role dynamics, prompting reallocation and adaptation of human responsibilities. Our findings contribute to the emerging field of Human-AI collaboration by exploring early-stage interactions and behavior towards AI in team settings. The study underscores the importance of role clarity and critical engagement for maximizing AI’s potential, and aims to contribute to building a continuum of understanding of Human-AI teaming.