UK researchers study AI’s potential to strengthen math teaching

LEXINGTON, Ky. (Sept. 18, 2026) — Helping students learn math requires teachers to tell students more than whether an answer is right or wrong. They need to be able to notice how students are thinking about the math problem. A new collaborative project between the University of Kentucky’s College of Education and Rowan University’s College of Education aims to strengthen teachers’ development of that skill, with the potential to ultimately improve mathematics student instruction. 

Supported by a $489,316 U.S. National Science Foundation award, researcher Jonathan Thomas, Ed.D., will lead a team exploring whether generative artificial intelligence (AI) can help evaluate teachers’ ability to recognize and respond to students’ mathematical thinking. This skill for teachers is known as “professional noticing.” 

UK’s principal investigator on the award, Thomas is professor and chair of the UK College of Education Department of Teaching and Learning, with co-principal investigators Cindy Jong, Ph.D., professor, and Robin Jocius, Ph.D., associate professor. Rowan University’s principal investigator is Molly Fisher, Ph.D., a professor in Rowan’s Content Area Teacher Education Department.  

“When it comes to helping students understand mathematics, professional noticing is a fundamental skill for teachers,” Thomas said. “However, it takes considerable time and expertise to evaluate a teacher’s professional noticing skills and provide useful feedback to help them grow in this area. That’s why we are exploring the use of generative AI. If it can be shown to evaluate teachers’ responses reliably, it could make this kind of assessment much more practical.” 

Aspiring teachers begin developing professional noticing skills during teacher preparation programs in college and may continue strengthening them throughout their careers. Teachers may further their professional noticing skills through working with a mentor or taking professional development courses. This study will use both groups — college students in a teacher preparation program and teachers already in the field — to examine how well generative AI can evaluate their professional noticing skills. 

“This study will directly impact Kentucky elementary teachers who participate, as we hope it will give them clear and rapid feedback on their noticing skills,” Thomas said. “That’s something that, before now, would have been impossible to do.” 

Providing teachers with evaluations of their professional noticing skills can help them see what they may have missed in a student’s thinking. With that feedback, teachers can strengthen their teaching skills and make better informed instructional decisions. 

“The goal is to give more teachers meaningful information about what they are noticing in students’ mathematical thinking and where there may be room to grow,” Jong said. “If AI can help us do that efficiently and reliably, it could expand access to feedback that is difficult to provide at scale.” 

The research team will start by teaching the AI what good professional noticing looks like.  

“We will give AI detailed instructions about what to look for in a teacher’s response and how to evaluate it. We’ll then use responses already scored by human experts to refine those instructions,” Jocius said. 

The researchers will then compare the AI’s evaluations with those made by trained human experts to see how closely they agree.   

“We want to see if the AI produces consistent results across different situations and ultimately determine whether it could be used as a practical way to evaluate professional noticing on a larger scale,” Fisher said.   

If successful, the research could make it easier for teacher preparation programs and schools to provide teachers with timely feedback on their professional noticing skills.  

The study will also help researchers better understand where generative AI can, and cannot, be trusted to evaluate complex teaching practices, providing guidance for its future use in teacher preparation and professional learning.   

“This study reflects the kind of innovative thinking our researchers bring to some of education’s most important challenges,” said UK College of Education Dean Nick Pace, Ed.D. “By exploring new ways to better understand and support effective teaching, their work has the potential to strengthen how teachers are prepared and make a meaningful difference for students across Kentucky and beyond.” 

This material is based upon work supported by the U.S. National Science Foundation under Award Number 2621046. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the U.S. National Science Foundation.