Generative AI in Collaborative Learning: A Systematic Review

• In Event: Human–AI Collaboration in Learning: Dialogue, Inquiry, and Facilitation Patterns

Abstract

This systematic review examines research trends, pedagogical supports, and reported learning effects of GenAI in collaborative learning. An analysis of 26 empirical studies indicates that all studies were published in 2024-2025, concentrated in higher education and engineering field, and largely utilized ChatGPT. Pedagogical supports were categorized into four themes: collaborative interaction, knowledge construction, metacognitive and regulatory, and task orchestration and procedural. Reported learning effects demonstrates predominance of individual cognitive gains and task efficiency over collective gains. While GenAI enhances task efficiency by streamlining coordination, it may reduce peer-to-peer engagement, potentially limiting the development of socially shared regulation and collective efficacy. This review concludes that GenAI should be designed to foster peer collaboration rather than merely function as an automated tool.

Keywords: Generative AI (GenAI); Computer-Supported Collaborative Learning; AI-supported Collaborative Learning; Systematic Review

Authors

• Sunyoung Kim, Korea University

• Jiyoon Jung, Korea University

• Insook Han, Korea University