Socratic: Enhancing Human Teamwork via AI-enabled Coaching

TL;DR

This paper tackles the challenge of limited access to human coaches in life-critical teamwork by introducing Socratic, an AI-enabled task-time coach. It formalizes team interaction as a Dec-POMDP and learns a latent-intent driven team model via multi-agent imitation learning (BTIL), then uses a TIC-based framework to generate context-aware interventions during task execution. Two human-subject studies (Movers and Flood) show that Socratic improves team performance with relatively few interventions and is perceived as helpful and trustworthy by participants. The results suggest that AI coaches can effectively augment human teams in high-stakes settings, with implications for scalable training and real-time support in domains like healthcare and disaster response.

Abstract

Coaches are vital for effective collaboration, but cost and resource constraints often limit their availability during real-world tasks. This limitation poses serious challenges in life-critical domains that rely on effective teamwork, such as healthcare and disaster response. To address this gap, we propose and realize an innovative application of AI: task-time team coaching. Specifically, we introduce Socratic, a novel AI system that complements human coaches by providing real-time guidance during task execution. Socratic monitors team behavior, detects misalignments in team members' shared understanding, and delivers automated interventions to improve team performance. We validated Socratic through two human subject experiments involving dyadic collaboration. The results demonstrate that the system significantly enhances team performance with minimal interventions. Participants also perceived Socratic as helpful and trustworthy, supporting its potential for adoption. Our findings also suggest promising directions both for AI research and its practical applications to enhance human teamwork.

Figures (11)

• Figure 1: Schematic of Socratic: an AI coach for enhancing teamwork during task execution. Blue arrows represent the workflow during the training phase, whereas black arrows indicate the workflow during the execution phase.
• Figure 2: Movers and Flood domains, detailed in \ref{['sec. domains']}. Team members can observe only the unshaded region of the environment.
• Figure 3: Snapshots of the Movers task from the first study (larger images are available in the appendix).
• Figure 4: Snapshots of Socratic's interactive user interface (larger images are available in the appendix).
• Figure 5: Team Scores: with and without Socratic.