Improving Teacher Training Through AI

School of Education Professor Jennifer Pankowski, EdD, specializes in working with neurodivergent students, embracing the role of technology in supporting both students and teachers in the classroom. Seidenberg Clinical Assistant Professor Tom Schmidt, DSc, has devoted his career thinking about the interplay between education and information systems, and how technology can cultivate a better learning experience.

Working together on a potentially groundbreaking research project was destined to happen.

“I connected with Jennifer, and I’ve so much enjoyed working her that it feels like we’ve always been in touch,” says Schmidt.

Both professors Pankowski and Schmidt have a strong interest in empowering neurodivergent students to reach their potential in the classroom. During the summer of 2024, Professor Schmidt continued his longstanding mentorship within this domain, serving as the faculty research advisor for then-Seidenberg student Ashley Peleg ’24; who based on her own experiences, sought to improve traditional teaching methods using XR technologies (virtual reality and augmented reality), so that teachers can better accommodate diverse learning styles in the classroom.

Fortuitously, this work aligned with a grant from the New York State Department of Education that Pankowski recently received. Entitled Enhancing Supports for Post Secondary Students with Disabilities, the grant focuses primarily on developing ways to help neurodivergent students have a better classroom learning experience—while additionally assessing if these innovations will improve the learning experience for all students.

As Schmidt elaborates, “By benefitting neurodivergent students, can we also further benefit everybody by the way we teach? This aligns very much with the research I’ve published over the past eight years, the science of successful learning. How do we adapt our teaching to make sure everyone feels included in the lesson? Can embracing (neuro)diversity make a better university for everyone?”

How do we adapt our teaching to make sure everyone feels included in the lesson? Can embracing (neuro)diversity make a better university for everyone?”

For many years, Pankowski has been employing avatars in the classroom to simulate complex situations that teachers may encounter in the classroom. Ashley’s research provided a foundation from which to further explore this area.

“Tom knew I worked with neurodivergent students and knew I worked with avatars. He reached out,” she says. “I had still been working with the avatar technology software I’d long been using, but through another platform, Edstutia, there was a lot more opportunity for innovation. He had me do a demo. That got us into this wormhole of—how do we really expand this thing?”

For Pankowski, a major focus of her practice is training her students to handle delicate and complex interpersonal situations with grace, poise, and expertise. Traditionally, this has been taught through case studies and role play simulations—both of which Pankowski note are not ideal, since case studies lack an experiential component, and role play can be tinged with implicit bias given the participants are already familiar with one another.

With Ashley’s research and the Edstutia platform as a launching point, Pankowski thought—what if, using this new technology, these simulations could be automated with artificial intelligence?

Thus, an idea was born. With direction from both Schmidt and Dan Buffone of Pace’s Learning Commons—who is also employing this technology—a few students are now helping to program the AI of Edstutia to automate a phone call simulation. The premise? A teacher is required to deliver not-so-good-news about a student…and must get the parent to come in for a parent-teacher conference before the parent hangs up the phone.

“We’re hoping by the start of the spring semester, our initial AI phone call will be ready to go,” notes Pankowski. “The goal is to get the parent (simulated by AI) to come in for a parent-teacher conference and not hang up the phone on you. If the AI-simulated parent hangs up the phone, the SOE student playing the role of the teacher will get a prompt to reach out to their professor, and then they’ll do the phone call with me.”

For Pankowski, this automation accomplishes several objectives simultaneously. For one, it significantly cuts down on the administrative load for faculty members tasked with creating simulations and role play scenarios, enabling her to focus on the actual skill-building and development of her students.

“I have a class of 30 education students. To really work one-on-one with students, or to have time to debrief in class, having an AI component that automates much of the process is hugely helpful. That frees up my availability, and now we’re able to bring the report into the classroom and discuss.”

To really work one-on-one with students, or to have time to debrief in class, having an AI component that automates much of the process is hugely helpful. That frees up my availability, and now we’re able to bring the report into the classroom and discuss.

Additionally, having a programmed AI bot removes some of the potential bias from the equation. It also offers the potential to significantly improve the learning experience for neurodivergent students, as it provides impartial analysis in a way that a standard role play simulation cannot.

“When we use this technology, we have a unique opportunity to be a fly on the wall—to give the feedback you’ll never hear from a parent. This allows us to discuss something like hey, this pinged on the AI as something you shouldn’t say, so let’s have a conversation; what were you trying to convey? And then we can have a redo.”

The redo component, Pankowski explains, is central to the value-add of this new technology. It provides Pace students ways to continuously develop real-world skills within the classroom setting—similar to the way a pilot in flight school builds their skills repeated simulations.

“The biggest benefit is that it’s giving that experiential component—the real-world application without the potential downfall of a real-world application,” explains Pankowski. “In the medical field, if you give someone bad advice, you don’t get a redo. In education, if you have a bad interaction with a parent, you don’t get a redo. What this does for students is allows them, like in gaming, to restart and get a do-over.”

In the medical field, if you give someone bad advice, you don’t get a redo. In education, if you have a bad interaction with a parent, you don’t get a redo. What this does for students is allows them, like in gaming, to restart and get a do-over.

As Pankowski and Schmidt look to further implement and expand this work, they hope to potentially be able to apply it other fields in which complex interpersonal interactions take place, such as within medicine or psychology.

As Pace is a leader in this field—Pankowski has noted that educators from Stanford, Michigan, Vanderbilt, Miami-Dade Schools, and as far as Israel have reached out to discuss this work—the potential for continued expansion and success can not only elevate the impactful work of our faculty and students, but make a difference in the lives of all students.

“No one is really looking at being a practitioner in this regard—we think about it for astronauts, racecar drivers, pilots. We don’t think about this for doctors, educators, psychologists, but they can benefit from this repeated simulation. That’s what we’re hoping to give, this Pace difference.”