Seoul National University Professor Han Jungwoo: "South Korea Must Add Materials Design Capabilities to Its Manufacturing Strength"
AI Accelerates Catalyst Discovery, but Commercialization Remains a Challenge...Downturn Leaves Little Room for Innovation

Professor Han Jungwoo of Seoul National University explains in an interview how artificial intelligence (AI) will transform scientific research and the development of new materials. Reporter Kim Jonghwa.

Professor Han Jungwoo of Seoul National University explains in an interview how artificial intelligence (AI) will transform scientific research and the development of new materials. Reporter Kim Jonghwa.

View original image


South Korea is a global manufacturing powerhouse. The production technologies it has developed in sectors such as semiconductors, batteries and petrochemicals are a key pillar of the country's competitiveness. But can manufacturing capabilities alone keep the country at the forefront of the global market in an era when artificial intelligence (AI) can design new materials?


Professor Han Jungwoo of the Department of Materials Science and Engineering at Seoul National University recognizes the strengths of South Korean manufacturing, but stresses that the country must build the capabilities to conceptualize and design technologies. Speaking with a reporter on October 30 at BEXCO in Busan after the closing of the Korean Society of Industrial and Engineering Chemistry's 2026 Fall Meeting, Professor Han said, "If the United States takes on the role of designing technologies, we may ultimately end up doing nothing but manufacturing."


He warned that South Korea's manufacturing competitiveness could also be limited if other countries take the lead in defining the concepts behind new technologies and products. He believes the country needs to combine AI-based materials design capabilities with its existing manufacturing strengths.


AI transforms materials development...Catalyst discovery system built in 13 days


Professor Han has researched computational science, which uses computers to predict the structures and properties of catalysts and materials. More recently, he has been using large language models (LLMs) and AI to search for new catalyst candidates. The approach moves away from methods that rely on researchers' experience and intuition, enabling AI to propose promising materials and perform the necessary calculations.


Professor Han's research team built a catalyst discovery system integrating AI and large language models (LLMs) in about 13 days. The system can search for candidate catalysts and return results within a few hours. However, this is a computer-based prediction stage; synthesizing the actual catalysts and verifying their performance still require separate processes.

AI-generated image

AI-generated image

View original image

The time required for individual calculations has also fallen sharply. Professor Han said, "In the past, it could take more than five days to run a single calculation and get a result, but now some calculations can be done in just a few seconds." This shows how computational techniques have advanced; it does not mean that the entire catalyst development process has become faster by the same degree.


Developing a new catalyst requires synthesizing candidate materials and measuring their performance. AI's advantage is not that it eliminates experiments, but that it makes the process of finding promising candidates more efficient.


Even if a material shows excellent performance in the laboratory, it cannot be put to use immediately in industry. Using it in a fuel cell requires the construction of cells and stacks, while applying it to a petrochemical process also requires the design of large-scale reactors.


Professor Han explained that "after confirming performance at the laboratory scale, the technology needs to be scaled up for use in actual industrial applications." He said his research team's current work focuses on the materials development stage. There is still a gap between the speed of AI-driven discovery and the pace of actual commercialization.

Large-scale petrochemical production facilities are located in the Yeosu National Industrial Complex in South Jeolla Province. As AI ushers in an era in which new materials and catalysts can be designed, voices are calling for South Korea's manufacturing industry to strengthen its research and development and technology design capabilities, moving beyond a focus on production. Provided by Yonhap News.

Large-scale petrochemical production facilities are located in the Yeosu National Industrial Complex in South Jeolla Province. As AI ushers in an era in which new materials and catalysts can be designed, voices are calling for South Korea's manufacturing industry to strengthen its research and development and technology design capabilities, moving beyond a focus on production. Provided by Yonhap News.

View original image

"Design in the United States, manufacturing in South Korea"...The next challenge for a manufacturing powerhouse


Professor Han is paying close attention to the entry of U.S. Big Tech companies into materials research. Meta is developing models to predict the properties of catalysts and materials through its AI-based atomic and materials simulation project, FAIR-Chem. Google DeepMind and Microsoft are also researching AI-enabled discovery and design of new materials, while Nvidia is involved in developing related technologies.


Professor Han said these companies see the materials and chemical sectors as future business opportunities. Companies equipped with AI technology, large-scale datasets and computing infrastructure are expanding their influence into the design of new materials.


He believes that South Korea could lead the process from developing new materials to manufacturing them by combining its manufacturing expertise with AI-based design capabilities. However, the country's research environment faces considerable constraints. Professor Han cited access to graphics processing units (GPUs) and reliance on overseas LLMs as major bottlenecks.


"One of the biggest bottlenecks in conducting research is GPUs," he said, explaining that domestic researchers have difficulty securing enough of the equipment they need. He added that as LLMs developed by overseas companies advance rapidly, reliance on those models is also increasing.

AI-generated images

AI-generated images

View original image

Technology advances, but companies face a downturn...Real-world constraints hinder innovation


Although AI technology is advancing rapidly, it is not easy for South Korean industries to make active use of it. Professor Han's assessment is that the prolonged downturn in the petrochemical industry is limiting the capacity to invest in new technologies.


"The chemical industry is in such a difficult situation that companies seem to have no room to focus on anything new," Professor Han said. "They are in a position where they need to restructure."


Before the COVID-19 pandemic, he also worked with companies on materials design and simulation. AI technology has advanced far beyond what it was then, but business conditions remain challenging for companies looking to adopt new technologies, he explained. Professor Han said the realities facing the industry should be understood rather than criticizing companies for their cautious approach.


Data security is another obstacle. Semiconductor, battery and chemical companies want to use AI in materials development, but are concerned that their core technologies and research data could leak outside their organizations. Sharing data among companies is also difficult.


Professor Han said companies recognize the need to use AI but are hesitant to adopt it actively because of data security concerns. The gap between large companies with sufficient in-house data and those without it could also become a problem.

Professor Han Jungwoo of Seoul National University discusses in an interview how artificial intelligence (AI) will transform scientific research and the development of new materials. Photo by Kim Jonghwa

Professor Han Jungwoo of Seoul National University discusses in an interview how artificial intelligence (AI) will transform scientific research and the development of new materials. Photo by Kim Jonghwa

View original image

Universities and research institutions also struggle to share data. In materials research, data formats vary depending on experimental conditions and measurement methods, so the data must be standardized before it can be used to train AI. However, organizing the data adds to researchers' workload. Professor Han acknowledged that even his own lab finds it difficult to systematically organize and share experimental data.


Professor Han also highlighted autonomous laboratories, where AI proposes materials and robots conduct experiments. This approach links AI predictions with robotic experiments and analysis of the results. However, because such facilities require expensive equipment and dedicated space, he proposed establishing facilities at government-funded research institutes or universities that multiple researchers could use jointly.


People still matter in the AI era...Research begins with "defining the problem"


How will the role of scientists change in an era when AI performs calculations and even proposes new materials? Professor Han said the rapid pace of technological advancement itself presents researchers with new challenges.


"Things are so different now from six months ago, and they have changed so much from a year ago that I can barely remember what it was like," he said. An algorithm developed over several months could lose its relevance with the arrival of a new LLM.

So what is the most important skill for scientists in the AI era? Professor Han's answer was "problem definition." He stressed, "You have to know what the problem is before AI can solve it. AI cannot define the problem for you. To define a problem, you need to study extensively and have a solid grounding in the scientific knowledge of your field."


Even when AI takes over calculations, the ability to decide what to research remains important.

[Science Scope] AI Designs, South Korea Just Manufactures? A Seoul National University Professor's Warning View original image

Professor Han also emphasized the importance of interaction among researchers. Conversations at academic conferences can be the starting point for new collaborative research. People may see professors having meals and socializing at conferences and think they are spending time on activities unrelated to research, but he said research discussions continue even over meals.


"Ultimately, everything we talk about while eating is research. All people think about is research," he said. For Professor Han, academic conferences are places to learn new information and build trust among researchers. Listening to other researchers' ideas and understanding the direction of the field can help shape future research plans.


"I think of every professor at another university who works in this field as a colleague," Professor Han said. Even if they are affiliated with different institutions, they share common concerns and work together.


Explaining why he attends academic conferences, he said, "You can sense what's happening in the field, and they give you plenty of opportunities to step back for a while and think about where we should go."



■ Professor Han Jungwoo of Seoul National University's Department of Materials Science and Engineering

He graduated from Seoul National University's School of Chemical and Biological Engineering in 2002, earned a master's degree in chemical engineering from Carnegie Mellon University in the United States in 2007, and received a doctorate in chemical and biomolecular engineering from the Georgia Institute of Technology in 2010. He later worked as a postdoctoral researcher at the Massachusetts Institute of Technology (MIT) in the United States. After serving as a professor at the University of Seoul and Pohang University of Science and Technology (POSTECH), he joined Seoul National University in 2023. His main research areas are the design of catalysts and energy materials using computational science and artificial intelligence (AI). He received the Ministry of Science and ICT's Scientist and Engineer of the Month Award in 2022 and the Korean Society of Industrial and Engineering Chemistry's Best Paper Award in 2024. Recently, he worked jointly with ETH Zurich in Switzerland to propose "Twin-loop," a strategy for designing gas-sensor materials that links AI simulations with experiments. The related perspective article was published in the international journal Nature Sensors in September 2026.


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing