Can AI Help Teachers Design Formative Assessments in Science Education? Evaluating the ADAPT-AI Tool
CAIRE Research Lab has received funding from the American Institutes for Research (AIR) to scale up and evaluate the ADAPT-AI system, a human–AI collaborative tool designed to support elementary science teachers in designing formative assessments.
The project will generate rigorous evidence on the validity and impact of ADAPT-AI in real classroom settings. Through a multi-phase evaluation involving elementary teachers and students in Washington State, the study will examine how AI-supported assessment design influences teacher confidence, student engagement, and academic outcomes.
The project includes two rapid-cycle evaluation phases:
- Phase 1 will involve disciplinary experts and fourth- and fifth-grade teachers to examine the quality and validity of ADAPT-AI learning tasks.
- Phase 2 will include approximately 80 elementary teachers and their students in a randomized study. Half of participating teachers will use ADAPT-AI+ immediately, while the other half will continue typical practice before later access. Teacher surveys, student surveys, and district data will be used to estimate impact.
This partnership with AIR marks an important milestone in advancing evidence-based, human-centered AI design for science education.
Learn more about the ADAPT-AI evaluation project.