Mapping Student-AI Interaction Dynamics in Multi-Agent Learning Environments: Supporting Personalised Learning and Reducing Performance Gaps
Project Overview
The document explores the integration of generative AI in education, emphasizing the use of multi-agent AI systems that emulate diverse instructional roles to create personalized and interactive learning experiences. It details how students engage with these systems, revealing patterns of interaction such as co-construction of knowledge and co-regulation. The study assesses the influence of these AI systems on learning outcomes and student motivation, highlighting their potential to bridge performance gaps and improve technology acceptance among learners. Overall, the findings suggest that generative AI can significantly enhance educational experiences by fostering deeper engagement and facilitating tailored learning pathways for students at various knowledge levels.
Key Applications
Multi-Agent AI System (MAIC)
Context: University students engaging with multiple AI agents in an online learning environment
Implementation: Students interact through a platform with various AI roles (teacher, peers) across modules on AI topics.
Outcomes: Increased learning gains, enhanced motivation, and improved technology acceptance, particularly for students with lower prior knowledge.
Challenges: Limited learning gains for students with high prior knowledge; potential ceiling effects and underutilization of AI support.
Implementation Barriers
Technological
Existing research focuses more on evaluating AI tools rather than understanding nuanced student-AI interactions.
Proposed Solutions: Future research should explore the interaction dynamics in multi-agent environments and develop guidelines for personalized system design.
Pedagogical
Students with higher prior knowledge may not benefit significantly from AI support due to ceiling effects.
Proposed Solutions: AI systems should adapt to different levels of prior knowledge, providing more reflective prompts for advanced learners.
Project Team
Zhanxin Hao
Researcher
Jie Cao
Researcher
Ruimiao Li
Researcher
Jifan Yu
Researcher
Zhiyuan Liu
Researcher
Yu Zhang
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Zhanxin Hao, Jie Cao, Ruimiao Li, Jifan Yu, Zhiyuan Liu, Yu Zhang
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