【摘要】 生成式人工智能驱动的对话系统(SDS)为二语口语能力发展提供了新的干预路径。本研究基于动态评估理念，探究教师和SDS协同干预模式下不同水平英语学习者口语产出的动态发展特征。通过追踪两名非英语专业学生(中高/中低水平)两周内10次教师—SDS协同干预训练中的表现，结合量化分析和质性数据，研究发现:学生英语口语能力得到整体性显著提升，不同水平学生英语口语语言特征发展呈现非线性轨迹且个体变异性突出。研究证实了“SDS自适应中介+教师诊断性介入”协同互补对学生英语口语能力发展有积极支持作用，为口语教学中人机协同的动态评估体系构建提供了实证证据。更多还原

【Abstract】 The spoken dialogue systems( SDS) driven by generative artificial intelligence has emerged as an innovative interventional tool for EFL learners’ oral proficiency development. Guided by dynamic assessment principles,this study constructed a teacher-SDS synergistic framework to examine the developmental trajectories of EFL learners’ oral output. Through a two-week case study experiment involving 10 speaking tasks with two Chinese EFL learners( intermediate-high vs. intermediate-low proficiency),triangulated data were collected for mixed methods analysis. Results revealed a statistically significant improvement in learners’ overall English oral performance,however,their developmental trajectories were nonlinear and exhibited marked individual variability. The findings confirmed the positive complementary effects of SDS adaptive mediation and teacher-led diagnostic intervention on oral proficiency development, offering empirical grounds for implementing dynamic assessment in the human-AI collaborative L2 speaking pedagogy.更多还原