Professor Chiu and Colleagues Awarded $5.5M to Support Teachers by Unlocking AI, Expanding Computer Modeling

In 2019, fifth and sixth grade students at a school in Charlottesville, VA were tasked with designing playgrounds that could effectively handle the water runoff from an adjacent hill, a task especially relevant for these students, as water and debris following a storm often limited their time outdoors. Creating computer models of the water runoff in their school yard, the students choose appropriate water-absorbing materials for their playground designs.

The lesson, now called the Water Resource Challenge, was a pilot project led by Jennifer Chiu, associate professor at the UVA School of Education and Human Development. And it was a success—the students expanded their understanding of science and engineering using the computational models. 

To expand the project, Chiu and colleagues from Vanderbilt University, Digital Promise, and SRI International created technologies that allowed students and teachers to build a computational model of their own school yard. Now, with a $3M National Science Foundation grant awarded across the partner institutions, they are expanding the project into schools in California, Tennessee, and in southwest Virginia, where water resources differ greatly.

“Some schools in California are in droughts, whereas other schools in Tennessee were built very close to flood plains,” said Chiu, who is the principal investigator on the study. “So, our project is trying to help teachers customize these technologies and projects to each school’s context.”

By working with 18 teachers and up to 4,800 students in these areas, Chiu and her team hope to build a generalized and adaptable model that any teacher anywhere can use.

Using AI to Support Computational Thinking in Science

With a second National Science Foundation award, Chiu and colleagues from North Carolina State University and Georgia Tech are working with teachers to leverage AI as a tool for computational lessons in science.  

The goal is to help augment teachers’ practice with a tool that can monitor where students are with a computational modeling or programming project. Through a dashboard, teachers will be able to provide hints, scaffolds, or feedback to help their students move along with their projects.

“AI can help teachers see things in their classrooms that they wouldn't be able to otherwise,” Chiu said. 

But AI can be intimidating. As part of this project, she and colleagues will involve teachers in co-designing the AI tools themselves, working to “peer into the black box of AI.”

“We want to help teachers be able to program AI tools themselves so that they're not as unfamiliar or scared about using them,” Chiu said. “Instead, they can open the black box, tweak it themselves and see, oh, they're not as uncomfortable or unfamiliar with it.”

Ultimately for Chiu, it is important for teachers to be in the loop as these tools are built so that they not only have relevance and utility in real classrooms but also so that the teacher stays in the loop and can both learn from and verify information provided to students.

A portion of the larger $2.5M grant awarded to NC State, Chiu’s $830,000 sub-award will fund three years of work with middle school teachers as they co-design and develop the AI tool, TRACES, with Virginia science teachers.