Strategies for Mitigating Racial Bias in AI Educational Technologies: A Literature Review (Poster 6)

• In Event: Artificial Intelligence, Gender, Race, Culture, and Environmental Justice in Education: An Intersectional Analysis of Bias, Equity, and Sustainability Globally

Abstract

The rapid advancement of artificial intelligence (AI) in the educational sector holds significant promise for enhancing learning experiences through personalized education, real-time feedback, and increased student engagement. However, AI deployment in education raises critical concerns about bias and inequality, particularly along lines of race and gender. This paper explores strategies to mitigate these biases to ensure AI- driven educational technologies promote equity and inclusivity.
The primary objective of this study is to identify effective strategies for mitigating biases in AI educational technologies. This entails analyzing the sources of bias in AI, assessing their impact on educational outcomes, and developing practical solutions to promote equity. By fostering fair, transparent, and inclusive AI systems, this research seeks to ensure that all students—regardless of background—benefit equitably from AI-driven education. Grounded in ethical AI development and inclusive education frameworks, this study integrates insights from critical race theory, white supremacy analysis, culturally relevant pedagogy, and other relevant racial theories to address systemic inequities in AI-driven learning environments. It analyzes how biases manifest in AI systems and their implications for educational equity, incorporating insights from machine learning and human-computer interaction to propose technical and policy-based interventions.
This poster draws upon interdisciplinary research from STEM fields and educational studies to discusses the emerging strategies in mitigating racial biases in AI development practices, and approaches to bridge the gap among researchers, policymakers and practitioners to prioritize equity in AI development thus to promote social justice through equitable and inclusive large language models across diverse fields.

Author

• Qiutong Ye, University of Wisconsin - Madison