Date
Publisher
arXiv
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 stereotype 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 (CEAT) with a prompt-engineered word extraction method within a Retrieval-Augmented Generation (RAG) 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.
Study at a glance
- Question: Can an automated tool reliably detect biases and stereotypes in AI-generated educational materials?
- Setting and data: AI-generated texts used in tutor training lessons.
- Method: Automated bias assessment evaluated against human-curated word lists (Contextualized Embedding Association Test, CEAT).
Why it matters
Biases embedded in AI-generated educational content, such as gender, racial, or national stereotypes, raise significant ethical and educational concerns. However, systematic methods for detecting and evaluating these biases in educational materials remain limited despite generative AI's growing use.
Key findings
- In AI-generated tutor training lessons, an automated bias tool matched human-selected test words with a near-perfect alignment score of 0.993.
- Automated bias screening provided reliable and consistent assessments of educational text.
- The automated method reduces human subjectivity and enhances fairness, scalability, and reproducibility in auditing AI-generated educational content.
Summary drafted with AI from the paper’s abstract and may contain errors. See the full paper for details.
What is the application?
Who is the user?
Who age?
Why use AI?
Study design
