���Ҵ�ý�ƽ������

Illustration by Hanna Nordwall

Researchers explore how AI will integrate with STEM education

What will classrooms look like in 20 or 30 years?

Will students build vocabulary through coaching from a smart speaker in the wall, getting feedback in real time? Will teachers consult with AI-powered software to understand where their class is falling behind and how to get back on track quickly? Maybe both? 

���Ҵ�ý�ƽ������ researchers are taking the first steps to develop the foundational theories, technologies and technical know-how needed to build these future classrooms — as well as the workforce needed to power them. 

Their research applies across grade levels and may change how future engineers learn about, engage with and teach STEM disciplines.

Sidney D’Mello is well positioned to talk about the dizzying potential and the roadblocks interdisciplinary researchers are addressing to get there. D’Mello is a professor in the Institute of Cognitive Science and in the Department of Computer Science, and he leads the National Science Foundation-funded AI Institute for Student-AI Teaming (iSAT).

Sidney D'Mello

“Current implementations of AI in the classroom have made great progress in supporting personalized learning with individual students, for example,” D’Mello said. “We are building off that type of work to start to envision the next-generation learning environment holistically as places where all students thrive in rich learning experiences co-created by students and teachers with the help of advanced technologies such as AI and machine learning.”

He added that inclusion and collaboration are a key focus for the center’s work. 

iSAT launched in 2020 as a partnership between university researchers and nearby school districts. In the first two years of operation, the institute has collected hours of classroom data, made advances in the integration of gesture and content analysis in teaching, and led workshops with high schoolers to help understand what students want and need from future AI partners.

“We have also made great strides in helping teachers implement our curriculum on AI literacy, reaching more than 2,500 middle school students in Colorado to date,” D’Mello said. “The institute has really proven to be a great way to bridge the 14 research areas from nine universities. The hope is that the end product of our multidisciplinary efforts will be greater than the sum of the individual parts contributed by each discipline.”

Many of the faculty involved with the institute are also part of the Engineering Education and AI-Augmented Learning Interdisciplinary Research Theme in the College of Engineering and Applied Science. Both groups are deeply invested in developing a diverse workforce of future researchers, administrators, policy officials and teachers in this space, said Co-Director Angela Bielefeldt.

Angela Bielefeldt

“The theme focuses specifically on education around STEM disciplines up through college versus more general primary school classrooms. But there are some large questions we are looking at as well that are applicable across this discussion,” said Bielefeldt, a civil engineering faculty member who also directs the Integrated Design Engineering Program. “One is the best way to design personalized educational materials developed by and along with AI. Another is how to best ensure equity, equality and inclusion for all students in these newly defined classroom spaces who are faced with new technologies.”

D’Mello said the AI institute will continue to organize activity in this research area across campus in the coming years. The current roster of researchers includes experts who understand how to create inclusive learning experiences that empower students with diverse identities, as well as those familiar with the latest research around the theories and frameworks needed to orchestrate student and teacher interactions with AI technologies.

“We also have researchers who can advance the foundational research needed for AI to 

be capable of monitoring, participating in, and facilitating collaborative learning conversations — all in the same place,” he said. 

“The center allows for a tremendous amount of knowledge sharing across all of those disciplines where ���Ҵ�ý�ƽ������ has both traditional strength and is well positioned to lead for the next 10 years.”