Human-AI Collaborative Learning: A Hybrid LLM Orchestration Framework Driven by Cognitive Dissonance
Exploring Science Academic Conference Series · 17.06.2026 · 24 · 14-19
DOI: 10.70267/iclpce.20261419 ↗Kurz gefasst
Fachartikel in der Zeitschrift Exploring Science Academic Conference Series, online erschienen am 17.06.2026. Verfasst von Yiqing Wu. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The rapid evolution of generative AI, particularly large language models (LLMs), is reshaping education from individualized instruction toward collective intelligence-driven collaborative learning. However, current intelligent educational systems face critical challenges, including fragmented collaboration ecosystems, isolated AI functionalities, and superficial learning analytics that hinder deep interactions among students, peers, and instructors. To address these gaps, this review proposes an integrated theoretical–technological–analytical framework grounded in constructivist learning theory and cognitive dissonance theory, conceptualizing knowledge as socially co-constructed. The framework comprises three core innovations: (1) a multi-agent collaborative architecture for real-time interaction; (2) a hybrid LLM orchestration mechanism for dynamic model deployment; and (3) a data-driven analytical engine leveraging semantic NLP to identify cognitive conflicts and trace collective intelligence trajectories. The framework advances next-generation intelligent collaborative learning systems by bridging learning sciences, artificial intelligence, and educational data mining. It also identifies key implementation challenges, including technical stability, ethical data governance, and pedagogical adoption, and proposes future directions such as affective computing integration, cross-institutional learning ecosystems, and multidimensional assessment reform. This work contributes a theoretically grounded and operationalizable design paradigm for AI-supported collaborative learning systems.
Zitieren
APA
Wu, Y. (2026). Human-AI Collaborative Learning: A Hybrid LLM Orchestration Framework Driven by Cognitive Dissonance. Exploring Science Academic Conference Series, 24, 14-19. https://doi.org/10.70267/iclpce.20261419
Vancouver
Wu Y. Human-AI Collaborative Learning: A Hybrid LLM Orchestration Framework Driven by Cognitive Dissonance. Exploring Science Academic Conference Series. 2026;24:14-19. doi:10.70267/iclpce.20261419
BibTeX
@article{Wu2026,
author = {Wu, Yiqing},
title = {Human-AI Collaborative Learning: A Hybrid LLM Orchestration Framework Driven by Cognitive Dissonance},
journal = {Exploring Science Academic Conference Series},
year = {2026},
volume = {24},
pages = {14--19},
doi = {10.70267/iclpce.20261419}
}
EUVIMED