Paper Type
Short
Paper Number
PACIS2026-1312
Description
As Artificial Intelligence evolves into autonomous agents, understanding how agent autonomy impacts team performance by reshaping social hierarchies is critical. This study conducted a code development field experiment spanning 160 team-weeks, combining mixed-effects modeling of varied autonomy modes with qualitative interaction log analysis. Results reveal a dual-pathway mechanism: heightened agent autonomy enhances performance by establishing expert power. Conversely, it elevates agent social status, inducing human blind compliance and reduced engagement, yielding a negative masking effect. Theoretically, this research disentangles expert power from social status, extending social hierarchy logic to human-AI hybrid teams. Practically, organizations must balance agents' professional empowerment with mitigating human cognitive inertia by embedding ultimate gatekeeping responsibilities into workflows.
Recommended Citation
Wang, Mingxu; Yao, Ruqing; Zhou, Jingmei; and Zhang, Jin, "From Passive Tools to Autonomous Teammates: How Agent Reshapes Team Performance through Social Hierarchy Restructuring" (2026). PACIS 2026 Proceedings. 5.
https://aisel.aisnet.org/pacis2026/ai_fow/ai_fow/5
From Passive Tools to Autonomous Teammates: How Agent Reshapes Team Performance through Social Hierarchy Restructuring
As Artificial Intelligence evolves into autonomous agents, understanding how agent autonomy impacts team performance by reshaping social hierarchies is critical. This study conducted a code development field experiment spanning 160 team-weeks, combining mixed-effects modeling of varied autonomy modes with qualitative interaction log analysis. Results reveal a dual-pathway mechanism: heightened agent autonomy enhances performance by establishing expert power. Conversely, it elevates agent social status, inducing human blind compliance and reduced engagement, yielding a negative masking effect. Theoretically, this research disentangles expert power from social status, extending social hierarchy logic to human-AI hybrid teams. Practically, organizations must balance agents' professional empowerment with mitigating human cognitive inertia by embedding ultimate gatekeeping responsibilities into workflows.
Comments
02-FutureofWork