【作 者】：
金牧兰， 钟绍春， 杨 澜， 张 璐
【关 键 词 】：
生成式人工智能； 数字教材； 教育大模型； 知识图谱； 数字人
【栏 目】：
其他
【中文摘要】：
数字教材作为数字技术与教育深度融合的产物，是推动教育数字化转型的关键要素。生成式人工智能的创新发展为数字教材建设提供了新方向与新路径。文章分析了生成式人工智能赋能数字教材的内涵与特征，从内容生成层、模态拓展层、交互处理层、路径推荐层和拟师助学层构建了生成式人工智能赋能数字教材建设的结构框架。在此基础上，提出了具体实践路径，即以教育大模型为基座，重塑多元化资源组织生成；以学科知识图谱为内核，支持智适应学习系统构建；以智能多向交互为拓展，促进泛在化深度知识创新；以多智能体为交互主体，实现个性化学习路径推荐；以虚拟数字人为代理，提供多场景定制化服务。通过数字教材的体系化重塑，助力数智时代拔尖创新人才培养。
【英文摘要】：
Digital textbooks, as products of the deep integration of digital technology and education, serve as a key element in promoting the digital transformation of education. The innovative development of Generative Artificial Intelligence (GenAI) has provided new directions and approaches for the development of digital textbooks. This paper analyzes the connotation and characteristics of GenAI empowering digital textbooks, and constructs a structural framework for GenAI-empowered digital textbook development from five layers: content generation layer, modality expansion layer, interaction processing layer, path recommendation layer, and teacher-like assistance layer. On this basis, specific practical pathways are proposed: leveraging educational large models as the base to reshape the organization and generation of diversified resources; taking disciplinary knowledge graphs as the core to support the construction of intelligent adaptive learning systems; taking intelligent multi-directional interaction as the extension to promote ubiquitous in-depth knowledge innovation; utilizing multi-agent systems as the interactive entities to realize personalized learning path recommendation; and employing virtual digital humans as agents to provide multi-scenario customized services. Through the systematic reconstruction of digital textbooks, we can facilitate the cultivation of top innovative talents in the digital- intelligent era.