生成式AI工具使用对英语专业学生写作焦虑与学习自主性的影响研究——基于自我决定理论的视角

The Impact of Generative AI Tools on English Majors’ Writing Anxiety and Learning Autonomy—From the Perspective of Self-Determination Theory

摘要: 基于自我决定理论，研究以120名英语专业本科生为对象，采用准实验设计，设置实验组(ChatGPT辅助写作)与对照组(传统写作)，通过量表收集数据并经SPSS 26.0分析。模拟结果显示：1) 实验组写作焦虑(M = 32.68, SD = 5.72)显著低于对照组(M = 41.35, SD = 6.21) (t = 7.83, p < 0.001)；2) 实验组学习自主性得分(M = 68.42, SD = 7.15)显著高于对照组(M = 56.79, SD = 8.33) (t = −8.56, p < 0.001)；3) 生成式AI通过满足自主(β = −0.32, p < 0.01)、胜任需求(β = −0.45, p < 0.001)降低焦虑，并通过满足自主(β = 0.38, p < 0.001)、归属需求(β = 0.29, p < 0.01)提升自主性。研究表明，合理使用生成式AI可缓解焦虑、提升自主性，为外语写作教学智能化改革提供依据。

Abstract: Based on Self-Determination Theory, this study conducted a quasi-experiment on 120 English majors, with an experimental group (ChatGPT-assisted writing) and a control group (traditional writing). Scale-collected data were analyzed via SPSS 26.0. Simulated results showed: 1) The experimental group had significantly lower writing anxiety (M = 32.68, SD = 5.72) than the control group (M = 41.35, SD = 6.21) (t = 7.83, p < 0.001); 2) Its learning autonomy score (M = 68.42, SD = 7.15) was significantly higher (M = 56.79, SD = 8.33) (t = −8.56, p < 0.001); 3) Generative AI reduced anxiety via autonomy (β = −0.32, p < 0.01) and competence needs (β = −0.45, p < 0.001), and enhanced autonomy through autonomy (β = 0.38, p < 0.001) and relatedness needs (β = 0.29, p < 0.01). Rational use of generative AI alleviates anxiety and improves autonomy, providing evidence for intelligent foreign language writing teaching reform.