
张晓轩:华中师范大学道德教育研究所2024级博士生

胡晓玲:华中师范大学道德教育研究所2024级硕士生

孙银光:华中师范大学教育学院副教授、道德教育研究所专职研究员
李璐:华中师范大学道德教育研究所2024级硕士生
邓诗弋:华中师范大学道德教育研究所2024级硕士生
陈晓雯:华中师范大学道德教育研究所2024级硕士生
Abstract: This study examines Chinese university students’ adoption of generative artificial intelligence (GenAI) tools by integrating the Theory of Planned Behavior (TPB), the Technology Acceptance Model (TAM), and AI literacy dimensions into a hybrid framework. Survey data from 1006 students across various majors and regions are analyzed using partial least squares structural equation modeling. Notably, AI literacy (i.e., students’ AI ethics, evaluation, and awareness) positively affect their attitudes, subjective norms, and perceived behavioral control, although the influence patterns vary according to the literacy dimension. Perceived privacy risks reduce AI trust, which mediates adoption behavior. Overall, core TPB pathways are validated, with behavioral intentions significantly predicting students’ actual use. Gender and regional differences moderate the key relationships. The results of this study suggest that enhancing students’ ethical and evaluative competencies, building user trust, and addressing privacy concerns could promote generative AI integration in education.
摘要:本研究将计划行为理论(TPB)、技术接受模型(TAM)与人工智能素养维度相整合,构建混合研究框架,探究中国大学生对生成式人工智能工具的采纳行为。研究采用偏最小二乘结构方程模型,对来自不同专业、不同地区的 1006 名大学生的调研数据展开分析。研究发现,人工智能素养(即学生的人工智能伦理认知、评估能力与应用意识)对其采纳态度、主观规范及感知行为控制均具有正向影响,且不同素养维度的作用机制存在差异。感知隐私风险会降低学生对人工智能技术的信任度,而信任在采纳行为中起到中介作用。总体而言,计划行为理论的核心作用路径得到验证,即行为意向能够显著正向预测学生对生成式人工智能工具的实际使用行为。此外,性别与地域差异对核心变量间的关系具有调节效应。本研究结论表明,提升学生的人工智能伦理判断与评估能力、构建用户信任体系、解决隐私安全顾虑,有助于推动生成式人工智能与教育教学的深度融合。
Keywords: generative artificial intelligence; AI literacy; technology acceptance; Theory of Planned Behavior; Technology Acceptance Model; privacy risk
关键词:生成式人工智能;人工智能素养;技术接受度;TPB模型;TAM模型;隐私风险
原文载于 Behavioral Sciences 2025年第15期





















