From Science Fiction to Reality: The Promise of Physical AI -

Published: November 2, 2026

Schools: Pratt School of Engineering

Scholars

Professor in the Department of Electrical and Computer Engineering
Associate Professor of Computer Science
Professor in the Thomas Lord Department of Mechanical Engineering & Materials Science

View Scholars@Duke Profile

Assistant Professor in the Thomas Lord Department of Mechanical Engineering and Materials Science
Dickinson Family Assistant Professor of Mechanical Engineering & Materials Science
Assistant Professor in the Department of Electrical and Computer Engineering

View Scholars@Duke Profile

John Cocke Distinguished Professor
Professor in the Department of Electrical and Computer Engineering
Director of the Institute of AI Engineering

View Scholars@Duke Profile

Sternberg Family Associate Professor of Electrical and Computer Engineering
Associate Professor of Electrical and Computer Engineering
Faculty Network Member of the Duke Institute for Brain Sciences

View Scholars@Duke Profile

The following is a summary of a series of stories appearing on Pratt School of Engineering.

These Pratt School of Engineering faculty members are studying how advanced systems are learning to navigate, understand and interact with the physical world.

Building in Security

One exciting possibility of physical AI is developing teams of autonomous vehicles and drones that can cooperate in real time. Imagine autonomous aircraft searching a disaster zone. Miroslav Pajic, a professor in the Department of Electrical and Computer Engineering, says the idea has great potential but also comes with vulnerabilities.

To avoid such attacks, Pajic says security should not be an afterthought; such features should be built in from the start.

Read the full story on Pratt School of Engineering.

What Robots Can Learn

Embedding AI into robots is changing how we interact with them. Instead of programming every behavior, “we can build robots that learn from their own experiences, develop internal models of themselves and their surroundings, and adapt when the world does not match what their designers expected,” says Boyuan Chen, Dickinson Family Assistant Professor of Mechanical Engineering and Materials Science.

His lab studies how a robot’s physical form shapes what it can perceive, learn, and do.

Lear more about robotics on Pratt School of Engineering.

Complex Decision-Making

For Yiran Chen, the John Cocke Distinguished Professor of Electrical and Computer Engineering, the challenge is finding ways to build systems that are efficient and powerful enough to run on resource-constrained devices while reliably supporting complex decision-making.

Efficiency, functionality, adaptability and trustworthiness are key components of physical AI systems, says Chen.

Physical AI systems must be able to see their surroundings and make decisions in real time, even when their cloud connection is unreliable.

Read more about Chen’s work on Pratt School of Engineering.

Safer Augmented Reality

What was considered science fiction just a couple of decades ago is now reality thanks to modern generative AI.

Maria Gorlatova, the Sternberg Family Associate Professor of Electrical and Computer Engineering, points to the 2008 Marvel film “Iron Man,” which she says “introduced one of the most sophisticated and intelligent examples of Physical AI embedded in an augmented reality platform” – J.A.R.V.I.S., an advanced artificial intelligence assistant created by the character Tony Stark.

She notes that while similar technology exists, it’s far less sophisticated. Her lab is leading pioneering research to make augmented reality technology safer and more reliable in high‑stakes settings.

Learn more about Gorlatova’s work on Pratt School of Engineering.

Source link