Science teachers are curious about generative AI, but might still working out how to move from generic uses toward integrations that genuinely support scientific practices. On April 17, 2026, the National GENIUS Center took that question into a room at the NSTA National Conference in Anaheim with a workshop titled GenAI to Enhance Science and Engineering Practices.
The session focused on one scientific practice in particular: developing and using models. Open-ended student drawings, limited class time, and the challenge of giving every student timely feedback make aredemanding for a teacher. The workshop tried to spark ideas how GenAI can help teachers meet those challenges.
Participants grounded their work in responsible and ethical use, sorting classroom scenarios into red-flag, yellow-flag, and green-flag categories. They drafted and tested their own scoring prompts on real student drawings for an NGSS particle-motion task, and tried ModelAgent, one of the GenAgents developed by the Center, which pairs a student drawing canvas with adaptive, practice-aligned support and a live teacher dashboard. Exit-ticket responses pointed to the prompt-design as particularly useful.
This workshop is part of the GENIUS Center’s broader effort to help science teachers move from generic AI use to content-specific, 3D learning-aligned integration and keeping teachers as the central decision-makers in their classrooms. We are thankful to the NSTA participants who shared their questions and expertise in the workshop, their feedback is already shaping the next round of PD design.
