大模型支持下高等数学人机协同课堂教学的构建与实施路径
Construction and Implementation Path of Human-AI Collaborative Classroom Teaching of Advanced Mathematics Supported by Large Language Models
摘要: 以大语言模型为代表的生成式人工智能正快速进入大学课堂。高等数学作为受众广、逻辑严密的公共基础课,既迎来个性化教学的机遇,也面临大模型数学推理错误、学生思维依赖等挑战。文章以“人机协同”为视角,借鉴张蜀青教授关于人工智能赋能数学课堂“自适应学习智能化升级、人机协同教学转向”的论述框架,并引入认知负荷理论与活动理论作为学理支撑,阐发大模型赋能高等数学教学的内在逻辑;构建“课前智能导学–课中双主协同–课后个性巩固”的全流程教学设计,并提出“准备–试点–常态”三阶段实施路径;进而结合国内外实证研究证据,分析该设计的可行性边界与主要风险。在此基础上,提出化错为源、梯度设限、以讲促学、教研赋能、分层参与、评价护航等优化策略,以期为同类课程改革提供参考。
Abstract: Generative artificial intelligence, represented by large language models, is rapidly entering university classrooms. As a public foundation course with a broad audience and rigorous logical structure, Advanced Mathematics is confronted with both the opportunity of personalized instruction and challenges such as the mathematical reasoning errors of large models and students’ cognitive dependence. From the perspective of “human-AI collaboration”, this paper draws on Professor Zhang Shuqing’s framework of “the intelligent upgrading of adaptive learning and the shift toward human-AI collaborative teaching”, and introduces cognitive load theory and activity theory to elaborate the intrinsic logic of empowering Advanced Mathematics teaching with large language models. It then constructs a whole-process instructional design characterized by “pre-class intelligent guided learning, in-class dual-subject collaboration, and after-class personalized consolidation”, and proposes a three-stage implementation path of “preparation-piloting-routinization”. Furthermore, drawing on empirical research evidence both in China and abroad, the paper analyzes the feasibility boundaries and major risks of the proposed design. On this basis, optimization strategies are put forward, including transforming AI errors into learning resources, imposing graded restrictions on AI use, promoting learning through student explanation, teaching through faculty research, providing differentiated participation pathways, and safeguarding implementation through process-oriented evaluation, with a view to providing a reference for the reform of similar courses.
文章引用:穆小焕. 大模型支持下高等数学人机协同课堂教学的构建与实施路径[J]. 教育进展, 2026, 16(11): 859-868. https://doi.org/10.12677/ae.2026.16112452

参考文献

[1] 中共中央 国务院印发 《教育强国建设规划纲要(2024—2035年)》[EB/OL].
https://www.gov.cn/gongbao/2025/issue_11846/202502/content_7002799.html, 2025-01-19.
[2] 徐宗本, 都琳, 黄廷祝. 人工智能时代的科学教育需要新定位和新范式[J]. 科学与社会, 2025, 15(1): 1-6, 14.
[3] 别敦荣. AI技术应用于大学教育教学的理论阐释[J]. 中国大学教学, 2024(5): 4-9, 2.
[4] 都琳, 徐爽, 徐宗本. 师-生-AI协同课堂: 人工智能赋能大学数学教育的载体及实践[J]. 中国大学教学, 2025(4): 59-65, 81.
[5] Sweller, J. (1988) Cognitive Load during Problem Solving: Effects on Learning. Cognitive Science, 12, 257-285.
https://doi.org/10.1207/s15516709cog1202_4
[6] Engeström, Y. (2001) Expansive Learning at Work: Toward an Activity Theoretical Reconceptualization. Journal of Education and Work, 14, 133-156.
https://doi.org/10.1080/13639080020028747
[7] 黄廷祝. 人工智能时代教学形态的主动变革[J]. 中国大学教学, 2025(1): 85-91, 107.
[8] Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö. and Mariman, R. (2025) Generative AI without Guardrails Can Harm Learning: Evidence from High School Mathematics. Proceedings of the National Academy of Sciences of the United States of America, 122, e2422633122.
https://doi.org/10.1073/pnas.2422633122
[9] 黄贤明, 李亚琼, 赵岩, 等. 生成式人工智能辅助初一学生解决数学问题的表现研究[J]. 数学教育学报, 2026, 35(3): 60-65, 92.
[10] 联合国教科文组织. 生成式人工智能教育与研究应用指南[R]. 巴黎: 联合国教科文组织, 2023.