NSF FF: Planning: A Teacher- and Guardian-Co-Designed System for Capturing Small-Group Scientific Argumentation in High-School Science

Award ID: 2627409

This FINDERS Foundry award explores whether an AIsupported and privacyaware tool can help teachers notice important but potentially "missed moments" during science learning. In many classrooms, students share ideas or get stuck, but educators are often not able to hear from every group or each learner at once. This project brings together an interdisciplinary team to co-create a privacy first system which identifies signs of confusion, strong ideas, or erroneous discussion by students to enable educators to provide support in real time. An interdisciplinary leadership team including educators and caregivers will shape how the tool is designed, used, what data can be kept, and how students can opt out. This collaborative effort builds trust while protecting student privacy. By giving teachers better insight into group conversations, the project supports stronger science learning while providing more learner voice, and a classroom environment where all students are supported. This FINDERS Foundry award advances the design of privacy-forward classroom AI tools that can help improve student learning and engagement in STEM. This project incorporates teachers' needs for capturing and processing student group audio; students' needs for acceptable and useful AI-extracted argumentation indicators; caregivers' needs for opt-out behaviors; and feasibility of on-device AI-driven audio processing. These considerations will result in a typology of points at which teacher presence would have changed student discussion; a functional specification of acceptable and unacceptable system behaviors, retention defaults, voice-identification policy, opt-out paths, and consent language acceptable to families; a feasibility assessment of on-device transcription; and an early design of a teacher-facing dashboard. By making privacy and teacher usability design requirements from the outset, the project offers an early, replicable design prototype for a classroom AI system that schools, teachers, and parents can trust, independent of any single set of indicators. 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