AI Could Speed Feedback for Math Teachers

Math and elementary teachers in training or in service may eventually receive faster feedback on their instruction if the study's AI tools prove accurate.

Updated on Nov. 9, 2026 in Financial Aid

Bold flat-color editorial illustration of a brass school desk bell resting centered on a plain cream cloth against a deep navy field. — The National Science Foundation has funded a two-year, $67,875 study by Rowan University's Molly Fisher testing whether AI can score how well math teachers respond to student thinking. AI Illustration. Upload story photo >

Live Poll

Would you trust AI to give your child's teacher useful feedback on classroom skills?

The National Science Foundation has funded a two-year, $67,875 study testing whether artificial intelligence tools can assess how well math teachers notice and respond to student thinking. The project, led by Rowan University professor Molly Fisher with a University of Kentucky team, will compare AI scores against expert raters' scores of responses from 150 in-service and preservice math teachers.

Why it matters

Teachers rarely receive feedback soon enough to use it in the moment, and evaluating professional noticing has proven challenging, which affects how quickly educators can improve their practice. For prospective and current teachers in education programs, faster and more reliable feedback could shape how their training is evaluated.

The two-year grant totals $67,875 and ends in July 2028. The study will include 150 in-service and preservice math teachers, with a planned recruitment of about 100 in-service elementary teachers and 50 preservice elementary teachers.

The players

National Science Foundation

The federal agency that funded the two-year, $67,875 study on AI-assisted assessment of teacher professional noticing.

Rowan University

The New Jersey public university where lead researcher Molly Fisher is a professor and leads the study.

University of Kentucky College of Education

The institution whose team is co-investigating the study alongside Rowan University.

The details

Researchers will train AI models on data from earlier studies so the tools can recognize examples of effective professional noticing, which is the ability to attend to, interpret and make instructional decisions based on students' thinking. Expert raters will score teachers' responses using validated protocols, and the AI tools, which include Claude, ChatGPT and a third tool still to be chosen, will score the same responses using refined prompts so the two sets of scores can be compared. No videos of students will be entered into the AI tools, and all teacher responses evaluated by AI will be de-identified.

Data collection with teachers is scheduled to begin in spring 2027.
The two-year grant ends in July 2028.

Education Landscape

Federally funded research into AI-assisted teacher evaluation reflects a broader push to find faster ways to give educators feedback than traditional expert scoring allows. The study sits within an ongoing effort to improve teacher preparation and in-service training across math and elementary education.

If you are a math or elementary education student or a practicing teacher, this study may eventually shape how teaching feedback is delivered, though data collection will not begin until spring 2027. Students interested in education programs can ask their institution's financial-aid office about grant-funded research opportunities that may be available to them.

The takeaway

Federally funded research like this one points toward faster, more reliable ways to give teachers feedback they can use in the moment. Education students who want to know whether research opportunities are available at their school can ask their financial-aid or education department office.