英语教育智能体：协作学习的设计与赋能

English Educational Agents: Design and Empowerment for Collaborative Learning

摘要: 生成式人工智能(GenAI)与大语言模型(LLMs)推动英语教育从CALL向ICALL深度变革，英语教育智能体(EEA)为缓解协作学习交互浅表化、反馈滞后等痛点提供技术解决方案。现有研究缺乏对英语学科特性(语言习得规律、跨文化需求)的针对性梳理。本研究采用系统性文献综述法(PRISMA 2020)与理论分析法，梳理四大数据库2014~2024年58篇高质量文献，构建“理论–技术–交互”三位一体设计框架，归纳五类核心角色及功能联动关系，解构三类交互模式运行机制。创新点在于厘清技术与教育理论适配逻辑，建立角色与交互模式动态联动机制，具体结合二语习得难点与英语学科特色技术实现路径，为人机协作英语教育提供兼具学理支撑与可操作性的实践路径。

Abstract: Generative Artificial Intelligence (GenAI) and Large Language Models (LLMs) are driving an in-depth transformation of English education from Computer-Assisted Language Learning (CALL) to Intelligent Computer-Assisted Language Learning (ICALL), and English Educational Agents (EEAs) offer a novel pathway to address the predicaments such as superficial interaction and delayed feedback in collaborative learning. Existing studies lack targeted elaboration on the disciplinary characteristics of English, including the laws of language acquisition and cross-cultural demands. Adopting the systematic literature review method (PRISMA 2020) and theoretical analysis, this study combs through 58 high-quality literatures published from 2014 to 2024 across four major databases, constructs a trinity design framework of “Theory-Technology-Interaction”, summarizes five categories of core roles and their functional coupling relationships, and deconstructs the operation mechanisms of three types of interaction modes. The innovation of this research lies in clarifying the adaptation logic between technology and educational theories, establishing a dynamic coupling mechanism between roles and interaction modes, thereby providing theoretical and practical support for human-machine collaborative English education.