Pop culture has often depicted robots as cold, metallic, and menacing, built for domination, not compassion. But at Georgia Tech, the future of robotics is softer, smarter, and designed to help.

“When people think of robots, they usually imagine something like The Terminator or RoboCop: big, rigid, and made of metal,” said Hong Yeo, the G.P. “Bud” Peterson and Valerie H. Peterson Professor in the George W. Woodruff School of Mechanical Engineering. “But what we’re developing is the opposite. These artificial muscles are soft, flexible, and responsive — more like human tissue than machine.”

Yeo’s latest study, published in Materials Horizons, explores AI-powered muscles made from lifelike materials paired with intelligent control systems. The technology learns from the body and adapts in real time, creating motion that feels natural, responsive, and safe enough to support recovery.
 

Muscles That Think, Materials That Feel

Traditional robotics relies on steel, wires, and motors, but rarely captures the nuances of human motion. Yeo’s research takes a different approach. He uses hierarchically structured fibers, which are flexible materials built in layers, much like muscle and tendon. They can sense, adapt, and even “remember” how they’ve moved before.

Yeo trains machine learning algorithms to adjust those pliable materials in real time with the right amount of force or flexibility for each task.

“These muscles don’t only respond to commands,” Yeo said. “They learn from experience. They can adapt and self-correct, which makes motion smoother and more natural.”

The result of that research is deeply human. For someone recovering from a stroke or limb loss, each deliberate movement rebuilds not just strength — it rebuilds confidence, independence, and a sense of self.

 

A Glove That Gives Freedom Back

One of the first real-world applications is a prosthetic glove powered by artificial muscles (published in ACS Nano, 2025), a device that behaves more like a helping hand than a mechanical tool. Traditional prosthetics rely on rigid motors and preset motions, but Yeo’s design mirrors the natural give-and-take of real muscle.

Inside the glove, thin layers of stretchable fibers and sensors contract, twist, and flex in sync with the wearer’s intent. The glove can fine-tune grip strength, reduce tremors, and respond instantly to the user’s movements, bringing dexterity back to everyday life.

That kind of precision matters most in the smallest tasks: fastening a button, lifting a glass, holding a child’s hand.

“These aren’t just movements,” Yeo said. “They’re freedoms.”

For Yeo, the idea of restoring freedom through movement has driven his research from the very beginning.
 

A Mission Rooted in Loss

Yeo's work is deeply personal. His path to biomedical engineering began with loss — the sudden death of his father while Yeo was still in college. That moment reshaped his sense of purpose, redirecting his focus from machines that move to technologies that heal.

“Initially, I was thinking about designing cars,” he said. “But after my father’s death, I kind of woke up. Maybe I could do something that helps save someone’s life.”

That purpose continues to guide his lab’s work today, building technologies that help people recover what they’ve lost.

Achieving that vision, however, means tackling some of engineering’s toughest challenges.
 

Soft Machines, Hard Problems

Creating lifelike muscles isn’t easy. They need to be soft but strong, responsive but safe. And they must avoid triggering the body’s immune system. That means building materials that can survive inside the body — and learn to belong there.

“We always think about not only function, but adaptability,” Yeo said. “If it’s going to be part of someone’s body, it has to work with them, not against them.”

His team calibrates these synthetic fibers like precision instruments — tested, adjusted, and re-tuned until they operate in sync with the body’s natural movements. Over time, they develop a kind of “muscle memory,” adapting fluidly to changing conditions. That dynamic adaptability, Yeo explained, is what separates a machine from a prosthetic that truly feels alive.
 

From Collaboration to Innovation

Solving problems this complex requires more than one discipline. It takes an entire ecosystem of collaboration. Yeo’s lab brings together experts in mechanical engineering, materials science, medicine, and computer science to design smarter, safer devices.

“You can’t solve this kind of problem in isolation,” he said. “We need all of it — polymers, artificial intelligence, biomechanics — working together.”

That collaborative model is supported by the National Science Foundation (NSF), the National Institutes of Health, and Georgia Tech’s Institute for Matter and Systems. In 2023, Yeo received a $3 million NSF grant to train the next generation of engineers building smart medical technology.

His team now works closely with healthcare providers and industry partners to bring these devices out of the lab and into patients’ lives.


The Future You Can Feel

The future of robotics, according to Yeo, won’t be defined by power or complexity but by feel.

“If it feels foreign, people won’t use it,” he said. “But if it feels like part of you, that’s when it can truly change lives.”

It’s the opposite of The Terminator, where machines replace us. Yeo is designing these machines to help us reclaim ourselves.

 

This story by Caitlin Hayes is shared jointly with the Cornell Chronicle newsroom.

Study co-author Joel E. Kostka is the Tom and Marie Patton Distinguished Professor and associate chair for Research in the School of Biological Sciences with a joint appointment in the School of Earth and Atmospheric Sciences. He also serves as faculty director of Georgia Tech for Georgia's Tomorrow. 

The Kostka Lab works in peatland ecosystems to quantify changes in microbial communities brought on by climate change drivers. In particular, next generation gene sequencing and omics approaches are employed to investigate the microbial groups that mediate organic matter degradation and the release of greenhouse gases.

Peatlands make up just 3% of the earth’s land surface but store more than 30% of the world’s soil carbon, preserving organic matter and sequestering its carbon for tens of thousands of years. A new study sounds the alarm that an extreme drought event could quadruple peatland carbon loss in a warming climate. 

In the study, published October 23 in Science, researchers find that, under conditions that mimic a future climate (with warmer temperatures and elevated carbon dioxide), extreme drought dramatically increases the release of carbon in peatlands by nearly three times. This means that droughts in future climate conditions could turn a valuable carbon sink into a carbon source, erasing between 90 and 250 years of carbon stores in a matter of months.

“As temperatures increase, drought events become more frequent and severe,  making peatlands more vulnerable than before,” said Yiqi Luo, senior author and the Liberty Hyde Bailey Professor in the School of Integrative Plant Science’s Soil and Crop Sciences Section, in the College of Agriculture and Life Sciences (CALS) at Cornell University. “We add new evidence to show that with peatlands, the stakes are high. We observed that these extreme drought events can wipe out hundreds of years of accumulated carbon, so this has a huge implication.”

“To me, this study is striking in that it shows that around 10 to 100 years of carbon uptake by one of the most important global soil carbon stores can be erased by just two months of extreme drought,” adds Joel Kostka, Tom and Marie Patton Distinguished Professor in Biological Sciences at Georgia Tech.

It was already well-established that drought reduces ecosystem productivity and increases carbon release in peatlands, but this study is the first to examine how that carbon loss is exacerbated as the planet warms and more carbon dioxide enters the atmosphere. The Intergovernmental Panel on Climate Change estimates extreme drought will become 1.7 to 7.2 times more likely in the near future. 

Read the full story in the Cornell newsroom. 

###

Other co-authors include Cornell postdoctoral researchers Jian Zhou and Ning Wei; senior research associate Lifen Jiang; and researchers from Georgia Institute of Technology, Florida State University, the U.S. Department of Agriculture (USDA), ETH Zurich, Northern Arizona University, the Australian National University, the University of Western Ontario and Duke University.

Funding for the study came in part from the National Science Foundation, USDA, the New York State Department of Environmental Conservation and the New York State Department of Agriculture and Markets.

The Center for 21st Century Universities (C21U) has announced the inaugural cohort of Bill Kent Family Foundation AI in Higher Education Faculty Fellows for 2025–26. This C21U-led fellowship program supports faculty projects that explore innovative, ethical, and impactful uses of artificial intelligence in teaching and learning. 

The fellows are Professor Flavio Fenton from the College of Sciences, Joy Arulraj from the College of Computing, Patrick Danahy from the College of Design, and Professor and Associate Chair of the School of Electrical and Computer Engineering Ying Zhang, from the College of Engineering. Each fellow will lead a project that advances AI’s role in higher education.

“We deeply appreciate the generosity of the Dr. Bill Kent family in establishing this first philanthropic gift to our new College. Their generous support will allow us to encourage practical applications of AI and foster an appreciation for its ethical use,” said William Gaudelli, inaugural dean of the Georgia Tech College of Lifetime Learning. “This Fellowship will ensure we grow and learn about its use thoughtfully, developing highly innovative and engaging pedagogical experiences for all life’s stages.”

Arulraj’s TokenSmith: Fast, Local, Citable LLM Tutoring introduces a privacy-conscious AI tutoring system for database courses that provides verifiable, course-aligned answers. Fenton’s AI as a Learning Assistant develops AI-enabled instructional modules for physics, neuroscience, and scientific writing to improve conceptual understanding and promote ethical AI use. Danahy’s AI-Enabled Design Ideation and Robotic 3D Printing with Open-Source Platforms integrates AI-driven design and robotic fabrication into architecture education while addressing ethics and sustainability. Zhang’s AI-Enabled Personalized Engineering Education expands personalized learning in large engineering courses through AI tutoring frameworks and integrates AI literacy into the curriculum.

“The Bill Kent Family Fellowship gives our faculty the resources and flexibility to experiment with AI in ways that directly benefit students and inform the future of higher education,” said Stephen Harmon, executive director of C21U.

The fellowship received 21 applications from all seven Georgia Tech colleges, reflecting the educational AI subject-matter experts for their units and the Institute as a whole. Fellows will develop and implement their projects during the 2025–26 academic year and share outcomes through C21U Learning Labs and other campus events.

The Bill Kent Family Foundation partnered with C21U to establish this fellowship and support faculty innovation at Georgia Tech. Through this program, the Foundation invests in projects that explore responsible and impactful uses of artificial intelligence in teaching and learning. By funding this initiative, the Foundation aims to empower educators to develop scalable instructional models, promote ethical AI practices, and prepare students for a future shaped by emerging technologies.