Carnegie Mellon University is shaping the future of physical AI—systems that can sense, react, and make decisions in the real world. Ahead of the AI Horizons Summit, Martial Hebert, dean of CMU’s School of Computer Science, explained why this field is becoming essential.
“Without intelligence, a physical system is just a machine,” Hebert said. Advances in vision and decision-making now allow robots, vehicles, and medical devices to operate with precision and safety, creating new opportunities across industries.
CMU has spent decades building this foundation. The university’s Robotics Institute, founded 45 years ago, focused on intelligent physical systems long before “physical AI” became a common term. Today, CMU researchers continue that work through projects like BrickGPT, which improves manufacturing through AI-driven decision-making.
The biggest technical hurdles, Hebert noted, are collecting large-scale real-world data and ensuring reliable performance in unpredictable environments. Researchers aim to develop trustworthy systems that can safely work alongside people.
With a strong base of startups, researchers, and tech talent, Pittsburgh is becoming a hub for physical AI innovation. CMU’s mix of research, entrepreneurship, and hands-on learning keeps its students and alumni at the center of this growing field.
“Whatever people learn today,” Hebert said, “there will be different approaches a few years from now. The key is to stay adaptable.”
Vraj Parikh
