[Paper Review] Generative AI Enhances Team Performance and Reduces Need for Traditional Teams

Ning Li, Huaikang Zhou+1 more|arXiv (Cornell University)|May 28, 2024

Big Data and Business IntelligenceBusiness, Management and Accounting3 citations

TL;DR

This study investigates generative AI's impact on team performance through a randomized controlled experiment with 435 participants across 122 teams. It finds that AI-augmented teams significantly outperform traditional human teams, with centralized AI use by few members yielding optimal results; however, individual-AI pairs still underperform compared to AI-integrated teams, indicating that while AI reduces reliance on traditional teams, full team integration maximizes performance gains.

ABSTRACT

Recent advancements in generative artificial intelligence (AI) have transformed collaborative work processes, yet the impact on team performance remains underexplored. Here we examine the role of generative AI in enhancing or replacing traditional team dynamics using a randomized controlled experiment with 435 participants across 122 teams. We show that teams augmented with generative AI significantly outperformed those relying solely on human collaboration across various performance measures. Interestingly, teams with multiple AIs did not exhibit further gains, indicating diminishing returns with increased AI integration. Our analysis suggests that centralized AI usage by a few team members is more effective than distributed engagement. Additionally, individual-AI pairs matched the performance of conventional teams, suggesting a reduced need for traditional team structures in some contexts. However, despite this capability, individual-AI pairs still fell short of the performance levels achieved by AI-assisted teams. These findings underscore that while generative AI can replace some traditional team functions, more comprehensively integrating AI within team structures provides superior benefits, enhancing overall effectiveness beyond individual efforts.

Motivation & Objective

• To examine how generative AI influences team performance in collaborative settings.
• To assess whether AI can replace or reduce the need for traditional team structures.
• To compare performance across human-only teams, individual-AI pairs, and AI-augmented teams.
• To investigate the effects of centralized versus distributed AI usage within teams.
• To determine whether multiple AIs per team yield incremental performance gains.

Proposed method

• Conducted a randomized controlled experiment with 435 participants forming 122 teams across various configurations.
• Assigned teams to conditions: human-only collaboration, individual-AI collaboration, and team-based AI augmentation.
• Used centralized AI deployment, where one or two team members interacted with AI on behalf of the group.
• Measured performance across multiple metrics, including task accuracy, speed, and quality of output.
• Compared performance outcomes across conditions using statistical analysis to identify significant differences.
• Evaluated the impact of multiple AIs per team to assess diminishing returns.

Experimental results

Research questions

• RQ1 Does generative AI enhance team performance compared to traditional human-only collaboration?
• RQ2 Can individual-AI pairs match or exceed the performance of conventional human teams?
• RQ3 What is the effect of centralized versus distributed AI usage on team outcomes?
• RQ4 Do teams with multiple AIs achieve further performance gains compared to single-AI teams?
• RQ5 To what extent can traditional team structures be reduced or replaced by AI integration?

Key findings

• AI-augmented teams significantly outperformed human-only teams across all performance measures.
• Individual-AI pairs matched the performance of conventional human teams but did not surpass them.
• Teams with multiple AIs showed no further performance gains, indicating diminishing returns.
• Centralized AI usage—where only a few members interacted with AI—yielded higher performance than distributed AI engagement.
• AI-augmented teams outperformed even the best individual-AI pairs, demonstrating the superiority of integrated team-AI collaboration.
• The findings suggest that while generative AI can reduce the need for traditional teams, full integration within team structures delivers the highest performance gains.