Promoting an Artificial Intelligence-Researcher-Teacher Network for Epistemic Practices in Science Education

Science education emphasizes engaging students with epistemic practices—disciplinary ways for students to think and work like scientists, such as creating models, explaining evidence, and making scientific arguments. However, designing these kinds of learning experiences can be challenging, and many teachers have limited opportunities for ongoing professional learning. While generative artificial intelligence (GenAI) can help teachers develop lessons and provide feedback to students, AI-generated learning activities often emphasize generic, procedural instruction rather than structured opportunities for engagement in epistemic practices. This project develops and studies an AI-assisted professional development program that supports high school science teachers to design and enact learning experiences that promote epistemic practices. The project includes two components: training and curriculum development. Teachers collaborate with an AI partner to co-develop lesson plans and co-analyze students’ learning artifacts. The project advances understanding of how AI can be responsibly designed to support teacher learning and improve science instruction. It engages 20 practicing educators and reaches approximately 2100 students in rural and suburban schools in Missouri and North Carolina. The project applies Design-Based Research, in which teachers and researchers collaborate to iteratively develop, evaluate, and refine the AI tool and professional learning materials. During the first year, five teachers co-design the tool and materials with researchers. In the second year, the tool is piloted with 10 teachers to evaluate usability and teacher learning. In the third year, the project expands to 20 teachers across two school districts. The research examines how teachers use AI to support lesson planning, student feedback, and science instruction, and how these approaches vary across teachers, science disciplines, and school contexts. The project develops case studies of teacher-AI collaboration from surveys of teachers’ AI literacy, empowerment to use AI for instruction, perceived tool usefulness, interviews with teachers and district coordinators, classroom observations, and students’ learning artifacts. The project aims to reveal specific patterns of teacher-AI collaboration to support epistemic practices. Technical contributions include (1) design principles to ground AI in education theories; (2) examination of AI’s capacities to provide feedback on multimodal student learning artifacts; (3) development of novel approaches to efficiently incorporate teacher feedback in technology design; and (4) systemic evaluation of multiple approaches to prompting and finetuning AI. This project was funded by the STEM K-12 program, which supports fundamental, applied, and translational research on STEM learning in both formal K-12 education and informal learning environments. 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.