摘要
生成式人工智能(Generative Artificial Intelligence)旨在利用人工智能技术自动化生成文本、图像、视频、音频等多模态数据,受到教育领域的广泛关注。其中,ChatGPT系统因其良好的自然语言理解和生成能力,体现出较高的多领域应用潜力。本研究以ChatGPT作为主要对象,基于其四项核心能力,即启发性内容生成能力、对话情境理解能力、序列任务执行能力和程序语言解析能力,探讨在教师教学、学习过程、教育评价、学业辅导四个方面的潜在教育应用。在此基础上,在真实系统中进行了习题生成、自动解题、辅助批阅等教育应用的初步验证。最后,本文进一步探讨了以ChatGPT为代表的生成式人工智能在教育应用中所面临的局限和对教育的启示。
Generative Artificial Intelligence aims to automate the generation of multimodal data such as text,image,video,and audio using artificial intelligence technology,and has received wide attention in education field.Among them,ChatGPT,a text content generation system,reflects high potentials for multi-domain applications due to its good natural language understanding and generation capabilities.In this study,ChatGPT is used as the main object to explore the potential educational applications in four areas of teacher teaching,learning process,education-al evaluation,and academic tutoring,based on its four core capabilities,namely,inspiring content generation capabil-ity,dialogue context understanding capability,sequential task execution capability,and programming language analy-sis capability.Based on this,preliminary validation of educational applications such as exercise generation,automatic problem solving,and assisted grading are conducted on the real system.Finally,the limitations faced by AIGC sys-tems represented by ChatGPT in educational application areas and the implications for education are further explored.
作者
卢宇
余京蕾
陈鹏鹤
李沐云
Yu Lu;Jinglei Yu;Penghe Chen;Muyun Li
出处
《中国远程教育》
CSSCI
2023年第4期24-31,51,共9页
Chinese Journal of Distance Education
基金
北京市教育科学“十四五”规划2021年度重点课题“人工智能驱动的新一代智能导学系统构建研究”(课题编号:CHAA21036)的研究成果。
关键词
生成式人工智能
ChatGPT
大语言模型
人工智能教育应用
Generative Artificial Intelligence
ChatGPT
Large Language Model
artificial intelligence applica-tion in education
作者简介
卢宇,北京师范大学教育学部未来教育高精尖创新中心副教授;余京蕾,北京师范大学教育学部教育技术学院博士研究生;通讯作者:陈鹏鹤,北京师范大学教育学部未来教育高精尖创新中心讲师,chenpenghe@bnu.eu.cn;李沐云,北京师范大学教育学部教育技术学院硕士研究生。
