Alex: A virtual peer that identifies student dialect

Our work in educational technology supports the deeper understanding of cultural phenomena that affect students in an educational environment. In this demo, we show a Virtual Peer system that supports collaborative science dialogue in two dialects: Mainstream American English and African-American Vernacular English, an often-stigmatized dialect. The most recently developed module in this system allows for the automatic classification of the students’ dialect in order to adjust the VP’s own dialect in response, and provide metalinguistic support. We are hopeful that this system will provide data to help close pervasive achievement gaps between students of different cultural backgrounds, while also providing important insights about the relationship between culture and learning that will impact educational technology design.