Using AI to Craft Linguistically Accessible Mathematical Tasks
Abstract
Creating mathematics story problems that maintain rigor while being linguistically accessible is a persistent challenge in today’s diverse classrooms. This study explores how 23 elementary preservice teachers (PSTs) used ChatGPT to adapt a fourth-grade fraction multiplication problem for multilingual learners (MLLs). By identifying both linguistic barriers and mathematical solution pathways, participants created differentiated versions aligned to various English proficiency levels. They developed visual word banks, simplified phrases, and structured supports while preserving the conceptual complexity of the mathematics. PSTs found ChatGPT helpful for generating language adaptations quickly but noted limitations with generated content and visuals. The experience demonstrated how thoughtful language modifications can support MLLs while maintaining high cognitive demand, positioning AI as a valuable tool for inclusive mathematics teaching rather than a replacement for teacher expertise.
Keywords: artificial intelligence, mathematics education, multilingual learners, task modification, elementary preservice teacher education
How to Cite:
Zhu, H., Bashirah, R. A., Eslami, S. M., Tabarak, A., Soto, M. M., Sutcliffe, K., Zhang, H. & Naranjo, C., (2025) “Using AI to Craft Linguistically Accessible Mathematical Tasks”, Ohio Journal of School Mathematics 101(1), 103-117. doi: https://doi.org/10.18061/ojsm.6617
Downloads
Download PDF
View
PDF
