|本期目录/Table of Contents|

[1]郭仕豪 朱德玲.赋能抑或危机:生成式人工智能使用对研究生科研能力的影响[J].清华大学教育研究,2025,(06):143-152.
　GUO Shi-hao ZHU De-ling.Empowerment or Crisis：The Impact of Generative AI Use on Graduate Students’ Research Capabilities[J].TSINGHUA JOURNAL OF EDUCATION,2025,(06):143-152.
点击复制

赋能抑或危机:生成式人工智能使用对研究生科研能力的影响
分享到：

清华大学教育研究[ISSN:1001-4519/CN:11-1610/G4]

卷:
期数:
2025年06期
页码:
143-152
栏目:
人工智能与教育
出版日期:
2025-12-20

文章信息/Info

Title:
Empowerment or Crisis：The Impact of Generative AI Use on Graduate Students’ Research Capabilities
作者:
郭仕豪 朱德玲
南京大学 教育研究院
Author(s):
GUO Shi-hao ZHU De-ling
Institute of Education, Nanjing University
关键词:
生成式人工智能; 研究生; 科研能力; 倒U型影响; 后人类主义
Keywords:
generative AI;  graduate student;  research capability;  inverted U-shaped effect;  posthumanism
分类号:
G643
文献标志码:
A
摘要:
在强人工智能时代的发展背景下,生成式人工智能使用对研究生的科研能力发展究竟是赋能还是危机,其影响机制有待被揭示。基于955份研究生调查数据进行实证分析,研究发现,对生成式人工智能的支持性使用与替代性使用对研究生科研能力有显著的倒U型影响,即适度使用能够提升科研能力,而过度使用则可能降低科研能力。此外,知识习得与知识启发对倒U型影响起到了显著正向的非线性中介效应,而研究生的人机关系认知在生成式人工智能使用与知识启发之间具有显著正向的调节效应,能够缓冲生成式人工智能使用造成的负面作用。研究结果对研究生个人适度使用生成式人工智能以及导师、培养单位引导研究生使用生成式人工智能助力科研具有一定的指导意义。
Abstract:
In an era marked by advances in strong artificial intelligence, the impact of generative AI on graduate students’ research capabilities is a pressing issue that warrants clarification. Based on an empirical analysis of survey data from 955 graduate students, this study reveals that both supportive and substitutive uses of generative AI have a significant inverted-U-shaped impact on research capabilities. Moderate utilization of generative AI can enhance research capabilities, while excessive utilization may reduce them. Additionally, knowledge acquisition and inspiration play significant positive but nonlinear mediating roles in this inverted U-shaped effect. Furthermore, graduate students’ perceptions of human-AI relationships significantly diminish the negative impact of generative AI on knowledge inspiration. The findings offer valuable guidance for graduate students’ moderate use of generative AI, as well as for mentors and institutions to help graduate students leverage it to promote scientific research.

相似文献/References:

[1]马 婧 姜 萌 刘欣欣.学术氛围促进研究生科研投入何以可能? ——认知需求、自我控制的作用[J].清华大学教育研究,2023,(06):93.
　MA Jing JIANG Meng LIU Xin-xin.The Effect of Academic Atmosphere on Graduate Research Engagement:The Role of Cognitive Needs and Self-Control[J].TSINGHUA JOURNAL OF EDUCATION,2023,(06):93.
[2]卢滇楠 党 漾 王宏宁 黎叙锐.生成式人工智能赋能高校课程教学:以“化工热力学”课程为例[J].清华大学教育研究,2024,(05):89.
　LU Dian-nan DANG Yang WANG Hong-yu LI Xu-rui.Empowering Higher Education Curriculum with Generative Artificial Intelligence: A Case Study of the “Chemical Thermodynamics” Course[J].TSINGHUA JOURNAL OF EDUCATION,2024,(06):89.
[3]让-加布里埃尔·加纳夏.高等教育中人工智能教育的若干挑战[J].清华大学教育研究,2025,(01):10.
　Jean-Gabriel Ganascia.Some Challenges in Teaching AI in Higher Education[J].TSINGHUA JOURNAL OF EDUCATION,2025,(06):10.
[4]袁培丽 宋 萑.空心的神像:生成式人工智能教育应用的审思[J].清华大学教育研究,2025,(04):19.
　YUAN Pei-li SONG Huan.The Hollow Idol: A Scrutiny of Generative Artificial Intelligence’s Application in Education[J].TSINGHUA JOURNAL OF EDUCATION,2025,(06):19.
[5]王传毅 强 浩 沈冬玲 刘 畅 周 钢.生成式人工智能如何辅助研究生知识生产 ——基于23万篇硕士学位论文的大模型语义识别研究[J].清华大学教育研究,2025,(04):71.
　WANG Chuan-yi QIANG Hao SHEN Dong-ling LIU Chang ZHOU Gang.How Does AIGC Empower Graduate-Level Knowledge Production ——A Semantic Analysis Based on 230,000 Master’s Theses Using Large Language Models[J].TSINGHUA JOURNAL OF EDUCATION,2025,(06):71.
[6]姚 荣 王 倩.生成式人工智能教育应用的风险及其法律规制[J].清华大学教育研究,2025,(06):131.
　YAO Rong WANG Qian.Legal Regulation of Risks in Generative Artificial Intelligence Educational Applications[J].TSINGHUA JOURNAL OF EDUCATION,2025,(06):131.

更新日期/Last Update: 2025-12-20