Racial Bias in AI Educational Tools (Poster 5)

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

Abstract

Contemporary societies undergoing rapid demographic transitions towards multilingualism and multiculturalism are facing an increasing utilization of artificial intelligence (AI) tools. In the field of education, AI tools are used in teaching and learning activities by people with multilingual and multicultural backgrounds. As these AI tools increasingly permeate various facets of daily life and education, there is a pressing need to scrutinize their potential implications within the framework of linguistic bias and racial bias.
This study aims to examine and analyze the intricate dynamics of language bias and inequality that exist in AI technologies, look into the literature to see how these biases are embedded in educational contexts and their influences, and find possible solutions to the field by answering the research questions: 1. What language bias and racial inequalities are presented in AI tools? 2. How are these biases embedded in educational contexts? 3. What are the potential solutions to create a more equal and culturally responsive educational environment with AI?
The first part of this study utilizes Critical Discourse Analysis (CDA) to investigate language bias and racial inequality in AI tools. Eight AI tools are analyzed to identify linguistic variation bias and propose strategies for mitigating bias and promoting linguistic equity. The second part synthesizes the literature studying AI related racial bias in educational contexts, from the perspectives of curriculum materials, student-AI interactions, and teachers’ AI tool use. The potential reasons and the bias and the feasible solutions are also investigated. The findings support the development of more effective and culturally responsive AI tools and the ways of using these tools in the education field.

Authors

• Xiangxiang Wang, University of Rochester

• Jialin Yan, University of Rochester