Korean ChatGPT KULLM, Turning Promise into Reality, Prof. Lim Heui-seok and the NLP&AI Lab
  • writing date 2025.03.06
  • author Communication Team
  • hits 80
Lim Heui-seok, Professor, Department of Computer Science and Engineering
Kang Myeong-hun (Integrated Master-Doctoral Degree Programs, 7th semester)
Koo Seon-min (Integrated Master-Doctoral Degree Programs, 7th semester)
Kim Min-hyeok (Integrated master-doctoral program, 2nd semester)
Kim Jin-seong (Integrated Master-Doctoral Degree Programs, 9th semester)
Park Chan-hee (Integrated Master-Doctoral Degree Programs, 1st semester)
Jang Yun-na (Integrated Master-Doctoral Degree Programs, Expected Graduate)
Korean ChatGPT KULLM, Turning Promise into Reality

임희석 교수와 NLP&AI 연구실 단체사진

With the increasing integration of ChatGPT and other AI-powered language models into various industries and everyday life, how advanced is Korea’s AI technology? To gain insight into the future, we visit a lab led by Professor Lim Heui-seok. The NLP & AI Lab, a leader in Korean AI research, developed KULLM (Korea University Large Language Model), a Korean counterpart to ChatGPT, marking a significant advancement in the development of Korean language models. The lab has also consistently presented original insights at EMNLP (Empirical Methods in Natural Language Processing), a top-tier conference in the field, with six papers presented in 2023 alone.


Professor Lim, the leader of the NLP & AI Lab, completed both his bachelor’s, master’s and doctoral degrees in Korea University’s Department of Computer Science and Engineering. He then began his teaching career at the College of Informatics, where he established the Human-Inspired AI Research lab. Through expanding the lab into a research institute, securing various support projects and industry-academia collaborations, and mentoring junior researchers, Professor Lim has come to value spontaneity and character over rigid performance expectations.

“Our lab submits papers to the top 10 conferences worldwide. Our students are naturally self-motivated, as we participate in the same conferences as Ivy League labs in the U.S. While talent is important, I prioritize character. My goal is for every student who passes through here to be a force for good and to excel in the field of AI.”

The ‘force for good’ that Professor Lim emphasizes is exemplified in the concept of human-facing AI, the core research direction of the Human-Inspired AI researchers and the NLP & AI Lab.

“Ultimately, I believe AI should benefit humanity. It should be designed as a tool to support and enhance human capabilities, not to replace them. KULLM, developed by the NLP & AI Lab, serves as a prime example of Human-facing AI.”

Our Lab Story

Due to the inherent nature of natural language processing, which requires a deep understanding of both computational science and linguistics, the NLP & AI Lab welcomes students from diverse academic backgrounds. Equipped with a robust infrastructure, including GPU resources that rank among the top three in Korea, and fostering an environment that encourages autonomous research, the lab helps researchers develop their capacity for engagement, patience, and strong collaborative skills.

Please tell us why you chose this lab.

Jin-seongI chose natural language processing (NLP) because I was fascinated by the idea of computers efficiently processing natural human language. My undergraduate major was in Spanish Language and Literature, and I believed that a linguistic perspective deriving from a literary background would be advantageous in NLP research. Although the computer science aspects were initially challenging, I made steady progress through dedicated effort, and I appreciate the many opportunities I’ve been afforded to utilize my second language.
Chan-hee My first experience with ChatGPT, during an interpreting session, revealed my comparative proficiency, prompting me to study NLP independently through resources like YouTube and Coursera. I recognized the increasing importance of technology and science in the field.
Seon-min Although my major was computer science, I wanted to explore a new field beyond what was currently being developed. At that time, AI was emerging, and I found natural language processing more compelling than image processing among the subfields involved.

What does a typical day look like in the lab?

Jin-seong Our workday usually starts with us arriving before 10 a.m. We check the number of members in each group and then work on our assignments and research. We typically go to lunch around 11 a.m. to avoid the cafeteria’s peak hours. In the afternoon, we focus on our research and assignments. If my professor is available, we often have dinner together. We usually leave work around 10 p.m., but it can be later depending on the research schedule.


임희석 교수

Professor Lim Heui-Seok

EMNLP 2024에 참석한 임 교수와 학생들

Prof. Lim and students at EMNLP 2024

What's your favorite fun memory of the lab?

Myeong-hunThe Abu Dhabi conference stands out as especially memorable, since travel to the Middle East is rare. We got to experience the desert, tour mosques, and create unforgettable memories.
Jin-seong We have this tradition of taking photos of each other at conferences. Myeong-hun is fantastic at landscape photography. When it comes to portraits, Sohn Soo-hyeon is our go-to, thanks to his older iPhone that’s known for its “Y2K” aesthetic (laughs).
Yun-naI have so many fond memories from conferences, especially with my teammates. It was great to spend the days diving into research and sessions, and then spend our evenings sightseeing and exploring. It was a rare chance to visit a country I probably wouldn't have gone to otherwise, so it was a nice bonus to my research experience.
Min-hyeokI also have a lot of great memories from lab dinners and gatherings. They were perfect opportunities to connect with both senior and junior members, especially at our regular spot, 'Memories of a Drink.

연구실에서 대화하는 임희석 교수와 학생들

What was your proudest moment in your research?

Seon-min I remember the achievement of publishing three papers with Kim Jin-seong at EMNLP. Publishing even one paper is tough, so preparing and presenting three at once was a major logistical challenge.

There must be some challenges when collaborating on a paper.

Jin-seong Collaboration definitely has its challenges. We have a talented team with diverse strengths—some excel at experimental work, while others are great at writing. But the key is effective teamwork, not just individual skills. For example, if one researcher prefers intense, short bursts of work and another likes a steady, day-by-day approach, their working styles can clash. However, in our lab, we have a lot of autonomy in choosing our collaborators. Occasionally, professors pair us with people we wouldn’t typically work with, which helps us develop the skills to navigate and overcome collaborative challenges.

Which MBTI types would be a good fit for the NLP & AI labs?

Jin-seong I believe ENFJ or INFJ types are particularly well-suited for our lab. We really prioritize teamwork and mutual encouragement, focusing more on collaborative effort than competitive outcomes. However, a balance between F (Feeling) and T (Thinking) is important for research.
Yun-na As an INTP, I find research very stimulating, though I’ll admit that interpersonal interactions can sometimes be a bit challenging for me.
Min-hyeok I’m an ENFP, and the lab’s dynamic, open research environment really fits my extroverted nature.

임 교수와 장윤나 학생

What do you do when your research isn’t successful?

Chan-hee I avoid trying to solve everything in one day. I keep a steady pace, and if I hit a roadblock, I address it the next day. Pushing too hard leads to burnout.
Myeong-hun My approach is to ensure continuous progress, no matter how small. I set a daily goal of writing three or four lines of my hopefully insightful thinking and ask for feedback to avoid tunnel vision. A simple suggestion from a knowledgeable colleague often helps me solve a problem.
Yun-na It is really important to refresh your thoughts. Constantly immersing yourself in papers from the same field can lead to stagnation. So, I explore research from unrelated areas or review papers by respected researchers. This helps spark new ideas and clears my mind.
Seon-min I use two strategies: if I’m feeling physically energized, I play a quick mobile game before getting back to work. But if I’m really stuck, I jot down my thoughts in a research journal and ask a nearby co-worker, ‘What do you think of this?’

Do you have any advice for students who want to join your lab?

Seon-min Don’t expect direct instruction in graduate school. It’s really a journey of self-discovery and independent learning.
Myeong-hun With a clear sense of purpose, you'll find that a wealth of resources is available to you.
Jin-seong Our lab values skills, emotional intelligence, and character, which foster a collaborative environment. But be prepared for a transformative experience—you may have more freedom when you enter than when you leave (laughs).
Min-hyeok Develop a researcher’s mindset, even as a student. I hope that the stories and images in this article inspire you to think, ‘I’m eager to join such a talented team!’
Yun-na Graduate school life is like a marathon. It requires endurance, persistence, and mental toughness, along with the ability to handle long hours. However, our lab also offers opportunities for international travel to conferences, where we get to enjoy diverse culinary experiences.

자료를 주고 받는 두 남학생

Graduates from the NLP & AI Lab, specializing in core AI technology development, have consistently excelled in both academia and industry, and they are leading the charge in AI innovation. As Professor Lim and his students confidently assert, the NLP & AI Lab, dedicated to fostering researchers who seamlessly blend rationality with empathy, will continue to nurture exceptional talent in the years ahead.


there is no attached file