Leveraging Bias in Computer Vision AI to Help Teachers learn to Create Bias-Free Learning Materials

• In Event: Critical Reflection of AI Integration in Education

Abstract

This paper reports research on teacher learning to create bias-free materials while learning about bias in computer vision AI. Presurvey, class artifacts, and artifact-based interview data were analyzed to address the overall research question, “How do preservice teachers learn to notice bias in ambient identity cues and create fair ambient identity cues for minortized students?” Findings are discussed in terms of participants’ noticing and criticizing bias in materials as well as their attempts to generate bias-free materials for minortized students.

Authors

• ChanMin Kim, Pennsylvania State University

• JungEun Ha, Pennsylvania State University

• Brian R. Belland, Pennsylvania State University