Scaling up AI-based professional development for math teachers everywhere

The United States has struggled to fill and retain math teacher jobs for years. In the 2011–2012 school year, 19% of public schools were unable to fill a teaching position for math, and by the 2020–2021 school year, 32% of public schools were unable to fill a teaching position for math. Compounding this math teacher shortage is the reality that many math teachers do not have positive experiences with the subject and do not receive adequate support. Fortunately, scalable and accessible online professional development (PD) for math instructors providing just-in-time feedback based on teachers’ understanding of content can improve the quality of math instruction. Just-in-time feedback provides teachers with training and development the moment they need it, instead of waiting for a PD trainer to assess and provide the feedback for the teacher. It also allows the teacher to apply new knowledge quickly.

Associate Professor Yasemin Copur-Gencturk’s newly published article positively reports the impact of artificial intelligence (AI)-powered PD on mathematics instruction with a prototype of a scalable and accessible program that creates an active learning environment for math teachers.

While evidence suggests that PD programs can improve teachers’ knowledge, instructional practices and learning outcomes for students, one problem that has not been addressed is whether effective programs are scalable so teachers in any location can access them. The barriers to accessing high-quality PD can be significant for teachers from rural and high-needs areas. Other factors such as lack of time and travel distances can limit teachers’ ability to participate in high-quality training.

Copur-Gencturk started her career as a math instructor in her native Turkey and understands the challenges that math teachers face. This on-the-ground experience inspired her research to examine alternative forms of learning environments, such as asynchronous online PD programs.

“While there has been online instruction in teacher education for some time, people have quit these programs because of the lack of interaction. We need to find a way to reach teachers anywhere and anytime to provide quality feedback, and this is where AI comes in,” said Copur-Gencturk.

To overcome the limitations of asynchronous PD programs by incorporating effective elements of in-person PD, Copur-Gencturk used intelligent tutoring systems to create an interactive, personalized learning environment by analyzing the performance of several instructors and interacting with them through a sequence of feedback cycles. The program that Copur-Gencturk and her research team developed utilizes a virtual facilitator that interacts with teachers through multiple activities. Math teachers from across the country worked through targeted content at their own pace, in their own space and on their own schedule.

Findings: The data reported in Copur-Gencturk’s new article shows that the teachers who participated in the AI-based PD program utilized mathematically richer tasks and created a more coherent and connected learning environment for students to build conceptual understandings than teachers who did not have access to the training. This piece builds off of Copur-Gencturk’s previous work over the past several years looking at how AI can be used to create effective, scalable teacher PD. The findings in this recently published article point to how AI-based PD programs that provide just-in-time feedback based on instructors’ understanding could improve the quality of their instruction. Tailored, just-in-time feedback from a virtual facilitator could be a substitute when human facilitators are unavailable. Reshaping how practitioners think about and conduct PD is one strategy that could be used to retain math instructors and support their continued development in the field.