DeepSeek在国内高校中的应用态势分析: 政策驱动、场景实践与风险应对  

Analysis of DeepSeek’s Application Trends in Chinese Higher Education Institutions:Policy,Practice and Risk

作　　者：马志强 崔鑫 尤欣雅 刘骥 苏福根[5] 郭桂真 Zhiqiang MA;Xin CUI;Xinya YOU;Ji LIU;Fugen SU;Guizhen GUO(School of Education,Jiangnan University,Wuxi 214122,Jiangsu;School of Design,Jiangnan University,Wuxi 214122,Jiangsu;Faculty of Education,Shaanxi Normal University,Xian 710062,Shaanxi;Research Division,Center for Education Management Information of the Ministry of Education,P.R.China,Beijing 100816)

机构地区：[1]江南大学教育学院,江苏无锡214122 [2]江南大学江苏“互联网+教育”研究基地,江苏无锡214122 [3]江南大学设计学院,江苏无锡214122 [4]陕西师范大学教育学部,陕西西安710062 [5]教育部教育管理信息中心研究处,北京100816

出　　处：《中国教育信息化》2025年第4期23-33,共11页Chinese Journal of ICT in Education

基　　金：2023年度无锡市“太湖人才计划”宣传文化人才项目资助计划社科理论类“无锡市智能化教育评价改革理论模型与实践路径研究”;2022年度中央高校基本科研业务费专项资金“协作知识建构会话智能化分析研究”(编号:JUSRP622048)。

摘　　要：DeepSeek凭借其卓越的内容生成、复杂的逻辑推理等性能,以开源模式与本地部署策略迅速渗入高等教育领域。在政策层面,中央与地方各级教育主管部门通过顶层设计明确了DeepSeek应用的价值导向与实践场景;高校自主制定行动计划,整合资源推动实践落地。在整体发展态势上,发达地区及“双一流”高校领先部署,理工类院校聚焦科研计算任务、师范类院校注重教学赋能、职教院校探索技能培养新模式等。在具体应用实践中,高校已将DeepSeek应用于教学、科研、管理与培训全流程,推动自主可控、深度适配的教学评一体化进程;促进科研效能提升、学科交叉攻关与产学研融通;实现稳定保障、多源集约的一站式服务;推进大模型教育课程建设与专项培训。然而,高校在应用DeepSeek方面也面临数据与隐私泄露、学术伦理与成果归属争议、教育主体认知退化、教育数字鸿沟加剧等风险。未来,应进一步加强统筹规划、校企合作、应用创新与风险防控,助力高等教育数字化转型向善、向好。With its excellent content generation and complex logical reasoning,DeepSeek has quickly penetrated into the higher education field with an open source model and local deployment strategy.At the policy level,education authorities at all levels clarified the value orientation and practical scenarios of DeepSeek application through top-level design while universities independently formulate action plans and integrate resources to promote practice.Development trends reveal leading adoption in developed regions and“Double First-Class”universities,with science and technology colleges focusing on scientific computing tasks,normal universities emphasizing pedagogical empowerment,and vocational colleges exploring new skills-training models.In practice,universities have applied DeepSeek to the whole process of teaching,scientific research,management and training to promote the integration of independent,controllable and in-depth adaptation of teaching evaluation;promoting the improvement of scientific research efficiency,interdisciplinary research and integration of industry-university-research;achieving stable guarantee,multi-source intensive one-stop service;promoting the construction of large model education courses and special training.However,challenges and risks persist,including data and privacy leakage,disputes over academic ethics and achievement ownership,cognitive regression among educators,and the deepening of the digital divide in education.In the future,overall planning,school-enterprise cooperation,application innovation,and risk prevention and control should be further strengthened to help the digital transformation of higher education become better and better.

关 键 词：DeepSeek 大模型 生成式人工智能 高等教育 人工智能教育 

分 类 号：G434[文化科学—教育学]