Many surveys show how quickly teachers are adopting generative AI, but far less is known about how they actually use it—whether it is mainly to save time or to improve learning—and how their institutional contexts shape those choices. This WestEd exploratory study looks closely at how math and science teachers use generative AI across instructional tasks, from planning and teaching to assessment and professional learning.

Drawing on 40 interviews with 15 STEM teachers and school and district leaders, the study finds that pedagogy, social context, and institutional context all shape how teachers use generative AI. It points toward clear but flexible policies, common tools that fit existing initiatives, AI narratives that promote innovation, content-specific and ongoing professional learning, a student-centered orientation, and special consideration for rural teachers. The study offers STEM educators; AI researchers; and state, district, and school leaders a context-grounded understanding to inform policy and practice.