使用CEAT框架对AI生成的教育内容进行自动化偏见评估
Automated Bias Assessment in AI-Generated Educational Content Using CEAT Framework
AI总结：
针对AI生成教育内容中存在偏见但缺乏系统检测方法的问题，本研究结合CEAT与提示工程词提取法，在检索增强生成框架下实现自动化偏见评估，在教师培训文本上取得r=0.993的高一致性，提升了审计的公平性与可扩展性。
摘要：
生成式人工智能（GenAI）的最新进展已变革了教育内容的创作方式，尤其是在辅导教师培训材料的开发方面。然而，AI生成内容中隐含的偏见——如性别、种族或国家刻板印象——引发了重大的伦理与教育层面的担忧。尽管GenAI的应用日益广泛，但用于检测和评估教育材料中此类偏见的系统性方法仍然有限。本研究提出了一种自动化偏见评估方法，该方法在检索增强生成（Retrieval-Augmented Generation）框架内，将Contextualized Embedding Association Test（语境化嵌入关联测试，即CEAT）与基于提示工程的词提取方法相结合。我们将该方法应用于辅导教师培训课程中使用的AI生成文本。结果显示，自动化词集与人工整理的词集之间具有高度一致性，皮尔逊相关系数r=0.993，表明该偏见评估方法可靠且稳定。我们的方法减少了人的主观性，提升了GenAI生成教育内容审计工作的公平性、可扩展性与可复现性。
英文摘要：
Recent advances in Generative Artificial Intelligence (GenAI) have transformed educational content creation, particularly in developing tutor training materials. However, biases embedded in AI-generated content--such as gender, racial, or national stereotypes--raise significant ethical and educational concerns. Despite the growing use of GenAI, systematic methods for detecting and evaluating such biases in educational materials remain limited. This study proposes an automated bias assessment approach that integrates the Contextualized Embedding Association Test with a prompt-engineered word extraction method within a Retrieval-Augmented Generation framework. We applied this method to AI-generated texts used in tutor training lessons. Results show a high alignment between the automated and manually curated word sets, with a Pearson correlation coefficient of r = 0.993, indicating reliable and consistent bias assessment. Our method reduces human subjectivity and enhances fairness, scalability, and reproducibility in auditing GenAI-produced educational content.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
