Google Research Opens Sign-Ups for Teacher-Requested AI STEM Simulation Pilot

Google Research has opened sign-ups for an upcoming pilot through which schools using Google Workspace for Education can request custom AI-generated STEM simulations shaped around a teacher’s curriculum and learning goals. It is also publishing a library of more than 30 English interactives that were generated by AI and reviewed and approved by teachers.

The release is a research experiment, not a general-purpose classroom chatbot. Google’s system turns a teacher’s proposed topic into a guided interactive simulation, beginning with learning objectives that the teacher can revise and approve. Those objectives steer the activity’s content and progression.

A simulation is meant to guide the work

Google describes generative UI as a way for AI models to construct an interface dynamically rather than rely on one coded in advance. Its intended output is a game-like learning experience with progressively harder levels, plus hints, formulas, feedback and worked solutions designed to guide students through a problem rather than simply present an answer.

Automated checks come before the classroom review

Google says it built iterative validation loops into generation rather than relying on a single pass. The system evaluates pedagogy, whether an activity can be solved, interface mechanics and visual quality. One evaluation opens a Chrome browser instance and interacts with the simulation, including attempts to push controls such as knobs to extreme values.

The public library covers physics, chemistry, biology, mathematics, computer science and earth science. Google says the released activities were AI-generated, then teacher-reviewed and approved. The library is available for public use, while the pilot gives eligible schools a route to request material tailored to their own STEM lessons.

Teacher feedback is encouraging, not an outcomes result

Google’s initial U.S. study involved 12 teachers, each requesting three custom interactives for their classrooms. The teachers gave the activities an average quality rating of 8 out of 10, according to Google. That measures perceived quality in a small initial group, not whether students learned more from the simulations.

The next test is in classrooms. Google says it plans user-experience research and field studies on student engagement and learning gains, leaving the educational impact of its guided-interactive approach unresolved.