【作 者】：
罗 恒， 廖小芳， 茹琦琦， 王志锋
【关 键 词 】：
生成式人工智能； 教师评语； 智能评测； 个性化教学； 初中数学
【栏 目】：
课程与教学
【中文摘要】：
教师评语是形成性评价分析的重要数据来源，但现有教师评语生成的质量在社会性、客观性和个性化方面存在一定的不足，生成式人工智能支持的教师评语为解决该问题提供了可能。生成式教师评语是指教师通过认知诊断技术得到的智能诊断数据与生成式人工智能平台进行交互，从而让生成式人工智能平台模拟教师社会性语言，生成数据驱动的个性化教师评语。在初中数学课堂中的实证研究发现：生成式教师评语有效地提高了学生的自我调节学习能力（Cohen's d=1.08， p=0.000）和学习动机（Cohen's d=0.59， p=0.003），对学生的深度访谈结果表明，生成式评语能作为考试的一种总结反思工具，体现了教师对学生个性化的关注和鼓励。研究结果为人工智能技术助力个性化、精准化教学提供了研究思路，为教育环境中持续评估和改进生成式人工智能技术的使用提供了建议。
【英文摘要】：
Teacher comments are an important source of data for formative assessment. However, the quality of existing teacher comments has certain deficiencies in terms of sociality, objectivity and personalization, and generative AI-supported teacher comments provide the possibility of solving these problems. Since intelligent diagnostic data obtained by teachers through cognitive diagnostic technology interacts with the generative AI platform, the generative AI platform can simulate teachers' social language and generate data-driven personalized teacher comments. An empirical study in a junior high school mathematics classrooms finds that generative teacher comments effectively improve students' self-regulatedlearning ability (Cohen's d=1.08, p=0.000) and learning motivation (Cohen's d=0.59, p=0.003). Through in-depth interviews with students, the results suggest that generative comments can be used as a summative and reflective tool for exams, reflecting the teacher's personalized attention and encouragement to students. The results provide research ideas for AI technology to facilitate personalized and precise teaching and learning, and provide recommendations for continuous evaluation and improvement of the use of generative AI technology in educational settings.