Authors: Melanie Kurimchak, Warren Li, and John Whitmer
https://osf.io/preprints/edarxiv/xjdnu_v3
Abstract
This paper synthesizes promising practices from a four-team research and development initiative aimed at improving AI performance in K-12 education. Drawing on projects focused on AI tutoring, teacher feedback, and benchmark development, we identify technical and partnership practices that increase the utility of AI systems for teaching and learning. Key findings address dataset design, evaluation frameworks, and collaboration structure, with consistent emphasis on pedagogical grounding over scale and clear division of expertise between education researchers and frontier model developers. These practices are offered as an evolving resource for the field as model capabilities and research infrastructure continue to advance.
Suggested Citation: Kurimchak, M., Li, W., & Whitmer, J. (2026, July 3). Not ready yet AI infrastructure EdTech market research. OSF. https://doi.org/10.35542/osf.io/xjdnu_v3