NSF FF: Planning: Earth Science AI Learning Foundry

Award ID: 2627487

This FINDERS Foundry award lowers barriers to high-quality instruction and improving Earth science outcomes for middle schoolers through an AI curriculum assistant is grounded in an AI architecture that generates responses based on a curated external knowledge base rather than solely on generalized training data. The knowledge base will be built from the ground up through a structured, statewide community curation process where resources will be vetted and validated by educators, families, and Earth science professionals. By grounding AI outputs in community-validated and expert reviewed Earth science content helps teachers trust the AI tool, improve science instruction, and prepare students for future STEM learning. The model is reproducible and transferable across academia, industry, and government for deploying more accurate and trustworthy AI tools for science education. This FINDERS Foundry award advances knowledge at the intersection of AI and science education by developing a novel approach to Retrieval-Augmented Generation systems built on stakeholder-vetted knowledge bases, creating a novel community-engaged educator-facing AI curriculum assistant. To generate the database, community resource submissions will be curated by assigning structured metadata: authority scores, standards alignment, grade and reading level, content type, and Earth sphere tags and then reviewed by master educators. The tool will generate Next Generation Science Standards-aligned lesson plans and instructional materials that save educators time and lowers barriers to high quality education for students. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Program Director: Lindsay Portnoy