摘要
人工智能教育成为当前国家数字化教育战略的重要着力点,人智信任更是人工智能赋能教育中最基础性的问题。虽然已有研究表明人智交互中信任影响最终的教学成效,但是人智信任机制却鲜有被关注。为此,研究首先聚焦人智交互现象,分析不同主体间的信任差异与特点,将人智信任解构为认知信任与情感信任双重作用过程,并针对性地设计了人智信任测度方法。通过对学生与ChatGPT的编程学习案例中的交互进行分析,表明认知信任对学业成绩有显著正向影响,而情感信任影响较弱;同时,任务难度的梯度变化以及时间的动态推移均会对学生与ChatGPT的交互信任产生影响。
Artificial intelligence education has become an important focus of the current national digital education strategy,and human-intelligence trust is the most fundamental issue in the empowerment of education by artificial intelligence.Although existing studies have shown that trust in human-intelligence interaction affects the final teaching effectiveness,the human-intelligence trust mechanism has received little attention.To this end,the research first focuses on the phenomenon of human-intelligence interaction,analyzes the trust differences and characteristics among different subjects,deconstructs human-intelligence trust into a dual action process of cognitive trust and emotional trust,and specifically designed a measurement method of human-intelligence trust.Through the interaction analysis in the programming learning cases between students and ChatGPT,it is indicated that cognitive trust has a significant positive impact on academic performance,while emotional trust has a relatively weak influence.Meanwhile,the gradient changes in task difficulty and the dynamic passage of time will both have an impact on the interaction and trust between students and ChatGPT.
作者
吴磊
刘宇欣
李珂新
WU Lei;LIU Yuxin;LI Kexin(Shandong Normal University,Jinan Shandong 250014)
出处
《现代远距离教育》
2025年第2期27-36,共10页
Modern Distance Education
基金
国家自然科学基金青年项目“知识建构社区中基于多模领域概念的知识创造路径识别与演化规律研究”(编号:62307026)
山东省高等学校青创科技支持计划(编号:2022RW090)
山东省社科联人文社会科学课题“数据驱动的知识建构社区知识扩散智能测度研究”(编号:2023-XWKZ-031)。
关键词
人智交互
信任机制
情感信任
认知信任
信任计算
Human-AI Interaction
Trust Mechanisms
Affectional Trust
Cognitive Trust
Trust Calculation
作者简介
吴磊,博士,山东师范大学教育学部副教授;刘宇欣,山东师范大学教育学部硕士研究生;通信作者:李珂新,山东师范大学教育学部硕士研究生。
