AI Boosts K‑12 STEM Learning: UNC‑Chapel Hill Joins $20 Million Funding Renewal

When you think about the future of school science labs, you might picture robots handing out test tubes. In reality, the next wave is being built in university labs, where researchers are teaching artificial‑intelligence systems to act like quiet, supportive tutors for kids.

Earlier this week the National Science Foundation announced it is extending the five‑year, $20 million award that funds the EngageAI Institute. The grant, originally launched at North Carolina State University, now runs through 2031 and carries an extra $20 million to keep the work rolling. UNC‑Chapel Hill is a core partner in the effort, alongside Vanderbilt, Indiana University and the nonprofit Digital Promise.

At the heart of UNC’s involvement is Professor Mohit Bansal, the John R. and Louise S. Parker Distinguished Professor of Computer Science. Bansal co‑leads the institute and heads its core AI research. He’s joined by distinguished professor Xiaoming Liu and assistant professor Roni Sengupta, both of whom bring deep expertise in computer vision, speech processing and the analysis of collaborative learning.

“The NSF’s continued backing signals that our work hits the sweet spot between cutting‑edge technology and real‑world teaching needs,” said Magnus Egerstedt, UNC’s executive vice chancellor and provost. “AI is reshaping everything, and we have a duty to make sure that transformation serves students and teachers alike.”

Since its 2021 debut, EngageAI has rolled out a handful of narrative‑driven learning environments where students solve problems together while an AI watches, listens and even interprets gestures. One of the flagship outputs is EngageVP, a framework that stitches together video, speech and motion data to give educators a window into how groups collaborate.

The next phase, set to kick off this summer, aims to go deeper. Researchers will build AI‑powered lessons that adapt on the fly, remembering a student’s progress across weeks or months and suggesting next‑step challenges tailored to individual strengths. Teachers, in turn, will get intuitive tools to weave 3‑D storytelling worlds into everyday lessons without needing to code.

“We’ve spent the last five years proving the technology works in the lab,” Bansal explained. “Now we’re turning that proof into classroom‑ready tools that can personalize STEM for every learner, not just a lucky few.”

Beyond the tech, UNC’s involvement showcases a broader commitment: using AI responsibly, testing its reliability, and ensuring it lifts society rather than replaces human insight. If all goes well, the next generation of K‑12 students could experience a science class that feels more like a collaborative adventure than a static lecture.

Keep an eye on future updates; the story is only just beginning, and the classrooms of tomorrow may already be shaping themselves today.