28. Mathmagician: Co-creating an AI-based Culturally Responsive Math Word Problem Generator with Educators and Learners

• In Event: Gallery Walk & Welcome Reception
  In Poster Session: Posters from Affiliates

Short Description

Equity pedagogies advocate that culture shall be a vehicle for learning. However, culturally responsive teaching is resource-demanding and difficult to implement in classrooms. This proposal introduces the design of Mathmagician, an artificial intelligence-based system that generates culturally responsive math word problems for 3-7th graders based on learners’ preferences. This study examines the co-design process of educators and learners’ involvement in creating training data, prototyping and prompting large language models to build this AI-based platform.

Authors

• Contributor: WENTING ZOU, The Pennsylvania State University

• Contributor: Zilong Pan, Lehigh University

• Contributor: Chenglu Li, University of Florida

• Contributor: Yuchen Yang, The Pennsylvania State University