Perception of an AI Teammate in an Embodied Control Task Affects Team Performance, Reflected in Human Teammates' Behaviors and Physiological Responses

2025-01-25

AI Summary

This study investigates how human perception of humanoid AI teammates affects team collaboration under embodied control tasks in virtual reality (VR), challenging the assumption that AI integration inherently enhances performance. Method: Leveraging a VR-based motor coordination task platform, we integrated multimodal physiological monitoring (electrodermal activity [EDA], heart rate variability [HRV]), behavioral coding, and speech interaction analysis. Contribution/Results: Human–humanoid AI teams exhibited significantly lower performance than human–human teams in physically demanding tasks, with performance deterioration intensifying as task difficulty increased. Humans displayed heightened physiological arousal (e.g., elevated EDA), reduced behavioral engagement, and diminished verbal communication—even when perceived AI trustworthiness increased. Critically, this work provides the first empirical evidence that enhanced AI “presence” can impair team dynamics rather than facilitate collaboration, thereby refuting the “AI integration = improved efficacy” heuristic. It establishes human physiological responses as a novel, essential dimension for evaluating human–AI teamwork efficacy.

Technology Category

• Humans and AI: Teamwork, Team formation
• Intelligent Robots: Embodied AI
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

The integration of artificial intelligence (AI) into human teams is widely expected to enhance performance and collaboration. However, our study reveals a striking and counterintuitive result: human-AI teams performed worse than human-only teams, especially when task difficulty increased. Using a virtual reality-based sensorimotor task, we observed that the inclusion of an active human-like AI teammate disrupted team dynamics, leading to elevated arousal, reduced engagement, and diminished communication intensity among human participants. These effects persisted even as the human teammates' perception of the AI teammate improved over time. These findings challenge prevailing assumptions about the benefits of AI in team settings and highlight the critical need for human-centered AI design to mitigate adverse physiological and behavioral impacts, ensuring more effective human-AI collaboration.

Problem

• Human-AI Team Performance
• Physical Activity Tasks
• Decreased Cooperation

Innovation

• Virtual Reality Gaming
• Human-AI Team Dynamics
• Enhanced AI Design for Collaboration