With $1.87M, Researchers Continue to Test How AI Can Support Special Education

UVA School of Education and Human Development Professor Michael Kennedy and his team have been awarded $1,875,000 from the U.S. Department of Education’s Office of Special Education Programs. This award comes on the heels of a $2.5M grant the team received from the same organization last fall.

Both projects aim to test how artificial intelligence can improve the teaching and learning of students with disabilities. With the new grant which will fund 100% of this new project, Kennedy and his colleagues will test how AI can support the creation of specific materials needed for teaching math and science vocabulary.

For five years, Kennedy and his team tested their evidence-based vocabulary practices with nearly 40 fifth through eighth grade science teachers. The study found significant positive impacts for thousands of students. But creating the materials is time consuming.

“We've been making materials by hand for years, and it takes a long time to make sure the content is accurate to state standards,” Kennedy said. “It's an intensive, expensive process. But what if AI is to the place where it could spit out quality materials that are in the neighborhood of what we've been doing by hand? Even if we had to tweak and modify them, that would be major.”

The team—including research assistant professors Rachel Kunemund and Olivia Coleman, and Christine Powell, assistant professor at Radford University—is working with Aidong Zhang, Thomas M. Linville Professor of computer science at UVA. 

“Zhang introduced us to AI that we didn’t know existed,” Kennedy said. “After showing us some prototypes of what she and her team were doing, we showed her what our slides look like and asked, ‘Can you make something like this?’ And they did.”

With this new funding, the team is expanding their work to include vocabulary practices for both math and science for teachers working with 3rd through 5th grade special education students and English language learners.

The research team will test if AI can produce the evidence-based materials in a way that supports teachers. They will also measure if it helps students make significant gains in mathematics and science learning.