An LLM-Guided Tutoring System for Social Skills Training
Michael Guevarra, Indronil Bhattacharjee, Srijita Das, Christabel Wayllace, Carrie Demmans Epp, Matthew E. Taylor, Alan Tay
TL;DR
The paper addresses the gap in social skills training (SST) where traditional classroom approaches fail to translate to real-world interactions. It introduces the GLOSS framework, an instructor-in-the-loop system that uses large language models to dynamically design realistic training scenarios, enable student rehearsal, provide immediate and delayed feedback, and visualize performance. Key contributions include a front-end narrative-graph builder, an interactive conversational simulator with real-time branching via LLMs, and an analytics tool to support instructor oversight and learner reflection. The framework emphasizes ease of scenario creation by instructors without programming, enabling personalized, domain-specific SST with flexible avenues for both scripted and open-ended interactions, and it supports web and VR interfaces to broaden accessibility and engagement.
Abstract
Social skills training targets behaviors necessary for success in social interactions. However, traditional classroom training for such skills is often insufficient to teach effective communication -- one-to-one interaction in real-world scenarios is preferred to lecture-style information delivery. This paper introduces a framework that allows instructors to collaborate with large language models to dynamically design realistic scenarios for students to communicate. Our framework uses these scenarios to enable student rehearsal, provide immediate feedback, and visualize performance for both students and instructors. Unlike traditional intelligent tutoring systems, instructors can easily co-create scenarios with a large language model without technical skills. Additionally, the system generates new scenario branches in real time when existing options do not fit the student's response.