Student engagement in collaborative learning with AI agents in an LLM-empowered learning environment: A cluster analysis

2025-03-03

AI Summary

This study addresses the challenge of characterizing heterogeneous patterns of student–AI agent collaboration in LLM-enhanced learning environments. Methodologically, it integrates hierarchical clustering with Epistemic Network Analysis (ENA) to analyze multimodal AI interaction logs, non-cognitive traits, and academic behavioral data from 110 undergraduate students. The analysis uncovers three empirically grounded learner archetypes: Active Questioners, Responsive Navigators, and Silent Listeners—distinguished by divergent interaction trajectories, cognitive engagement intensity, and network structure. Results quantitatively demonstrate significant differences in epistemic behaviors and collaborative dynamics across types, revealing distinct mechanisms of AI-mediated cognition and participation. The contribution is a theoretically grounded, interpretable, and actionable learner typology framework that supports fine-grained, real-time pedagogical interventions. This advances adaptive educational systems beyond one-size-fits-all AI integration toward genuinely personalized, co-constructive human–AI learning partnerships.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Agent/AI Theories and Architectures
• Humans and AI: Learning Human Values and Preferences

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Integrating LLM models into educational practice fosters personalized learning by accommodating the diverse behavioral patterns of different learner types. This study aims to explore these learner types within a novel interactive setting, providing a detailed analysis of their distinctive characteristics and interaction dynamics. The research involved 110 students from a university in China, who engaged with multiple LLM agents in an LLM-empowered learning environment, completing coursework across six modules. Data on the students' non-cognitive traits, course engagement, and AI interaction patterns were collected and analyzed. Using hierarchical cluster analysis, the students were classified into three distinct groups: active questioners, responsive navigators, and silent listeners. Epistemic network analysis was then applied to further delineate the interaction profiles and cognitive engagement of different types of learners. The findings underscore how different learner types engage with human-AI interactive learning and offer practical implications for the design of adaptive educational systems.

Problem

• Explores learner types in AI-enhanced education
• Analyzes interaction dynamics with LLM agents
• Classifies students into engagement-based clusters

Innovation

• Integrates LLM models for personalized learning
• Uses hierarchical cluster analysis for student classification
• Applies epistemic network analysis for interaction profiles