Skip to main content Skip to navigation

From MOOC to MAIC: Reshaping Online Teaching and Learning through LLM-driven Agents

Project Overview

The document explores the transformative impact of generative AI, particularly large language models (LLMs), on online education, highlighting the development of the Massive AI-empowered Course (MAIC). This innovative framework employs LLM-driven multi-agent systems to enhance scalability and adaptivity in learning environments, aiming to deliver personalized educational experiences and improve overall outcomes. Through intelligent tutoring systems and AI assistants, MAIC seeks to engage students more effectively and elevate perceptions of teaching quality. Preliminary experiments conducted at Tsinghua University involving over 500 students have shown positive results in terms of student engagement and satisfaction with educational quality. However, the document also acknowledges ongoing challenges related to further enhancing personalization and adaptability in AI applications. Overall, the integration of generative AI in education is positioned as a promising avenue for fostering individualized learning and improving educational effectiveness, despite certain hurdles that need to be addressed.

Key Applications

MAIC (Massive AI-empowered Course)

Context: Online education for diverse learners, particularly in higher education settings.

Implementation: Utilizes LLM-driven multi-agent systems to create personalized and adaptive learning experiences.

Outcomes: Improved student engagement, enhanced understanding of course objectives, and positive feedback on AI instructor quality.

Challenges: Personalization of learning experiences for individual student needs, maintaining engagement during passive learning modes.

Implementation Barriers

Technical

Challenges in integrating various AI technologies into a unified educational framework.

Proposed Solutions: Ongoing development and refinement of AI systems and workflows to ensure seamless operation.

Pedagogical

Difficulty in balancing the automation of teaching with the need for human interaction and emotional support.

Proposed Solutions: Incorporating human-in-the-loop design principles to maintain educator involvement in the learning process.

Ethical

Concerns regarding data privacy and potential biases in AI systems.

Proposed Solutions: Implementing strict data protection measures and fairness-focused auditing procedures.

Engagement

Students may prefer passive learning modes, limiting opportunities for active participation.

Proposed Solutions: Designing AI tools that actively encourage student interaction and inquiry.

Project Team

Jifan Yu

Researcher

Zheyuan Zhang

Researcher

Daniel Zhang-li

Researcher

Shangqing Tu

Researcher

Zhanxin Hao

Researcher

Rui Miao Li

Researcher

Haoxuan Li

Researcher

Yuanchun Wang

Researcher

Hanming Li

Researcher

Linlu Gong

Researcher

Jie Cao

Researcher

Jiayin Lin

Researcher

Jinchang Zhou

Researcher

Fei Qin

Researcher

Haohua Wang

Researcher

Jianxiao Jiang

Researcher

Lijun Deng

Researcher

Yisi Zhan

Researcher

Chaojun Xiao

Researcher

Xusheng Dai

Researcher

Xuan Yan

Researcher

Nianyi Lin

Researcher

Nan Zhang

Researcher

Ruixin Ni

Researcher

Yang Dang

Researcher

Lei Hou

Researcher

Yu Zhang

Researcher

Xu Han

Researcher

Manli Li

Researcher

Juanzi Li

Researcher

Zhiyuan Liu

Researcher

Huiqin Liu

Researcher

Maosong Sun

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Jifan Yu, Zheyuan Zhang, Daniel Zhang-li, Shangqing Tu, Zhanxin Hao, Rui Miao Li, Haoxuan Li, Yuanchun Wang, Hanming Li, Linlu Gong, Jie Cao, Jiayin Lin, Jinchang Zhou, Fei Qin, Haohua Wang, Jianxiao Jiang, Lijun Deng, Yisi Zhan, Chaojun Xiao, Xusheng Dai, Xuan Yan, Nianyi Lin, Nan Zhang, Ruixin Ni, Yang Dang, Lei Hou, Yu Zhang, Xu Han, Manli Li, Juanzi Li, Zhiyuan Liu, Huiqin Liu, Maosong Sun

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies