As the development of artificial intelligence (AI) technology is shaking the paradigm of industries worldwide, leading countries, including the United States and China, are engaged in a power competition. In particular, China, which has made substantial investments in AI and semiconductors as a national project, has recently sparked a DeepSeek craze, revealing its significant level gap with South Korea.
South Korea, which once prided itself as an IT powerhouse, has seen its AI research and development primarily grow within the institutional sector of corporations, and even at this point, expecting large-scale projects or budget support led by the government is difficult. The inadequacy is bluntly revealed when examining the state of AI research and development at Seoul National University, one of the country’s top universities.
Professor Yun Sung-ro, a key figure in South Korea's AI academic community and former chairperson of the Fourth Industrial Revolution Committee, told ChosunBiz on the 4th, "Even Seoul National University, one of the best universities in the country, has a server room for AI research that students refer to as 'Aoji coal mine' due to inadequate heat control," adding, "We do not have a budget to purchase cooling equipment to dissipate heat, and electricity usage is limited, making it difficult to conduct proper AI research and development."
He pointed out, "It is so poor that the researchers are struggling to cover the expense of purchasing graphics processing unit (GPU) equipment, and they have to study ways to develop AI solutions without GPUs. With such circumstances, not only is there a decline in technological advancement, but the number of talent to lead the future of the AI industry is also rapidly decreasing, and it seems that we have to provide an even lower level of education than 10 years ago."
Professor Yun emphasized that the treatment of the scientific field, which will be the foundation for the advancement of AI technology, must also change. He noted, "While there are pros and cons to government-led AI industry development, it is necessary for the government to step in, especially in areas where talent development and the institutional sector are lacking. In particular, in South Korea, the treatment of the scientific and technological field is drastically inferior compared to China. There are even claims among the elders that support is lacking compared to the times of former President Park Chung-hee, when scientists were treated exceptionally well and efforts were made towards technological advancement." Next is a Q&A with Professor Yun Sung-ro.
─You served as chairperson of the Fourth Industrial Revolution Committee in the previous government. What significance do you see in the long-term plans and investments led by the government for the AI industry development?
"I believe the right path is for the private sector to take the lead. However, when corporations are leading but face limitations in terms of strength or investment funds, and when the market itself is not formed, there are cases where they abandon self-development and adopt foreign AI models entirely, which is uncommon in our country. Ultimately, it is necessary to encourage corporate investment, but they hesitate because the return on investment is unclear. It must be structured in a way that allows for the utilization of AI, commercialization, and profit creation. Viewed in that light, the government's role is significant. It must act by providing extraordinary treatment for establishing research institutes and attracting outstanding talent."
─Do you think more proactive government efforts are needed regarding AI investment and talent development?
"Looking back at the era when the government led the advancement of science and technology, for instance, during the time of former President Park Chung-hee, exceptional treatment was given to scientists returning home. Of course, while the treatment of researchers at national research institutes has steadily improved, it is still lacking greatly compared to the long-term support provided by the U.S. and Chinese governments. This does not mean that the government should unconditionally pour in funds. The U.S. Stargate project was an initiative to induce private investment rather than direct government investment, and most of the funding was from private investments. This is because the government has a well-organized plan to induce and coordinate investments."
─One of the major driving forces behind the development of AI businesses in the U.S. and China is the revitalization of industry-academia cooperation. What about in South Korea?
"The conditions for academia to conduct AI research that can contribute to corporations are not favorable. This can be seen at Seoul National University. Although it is one of the representative universities in our country, the situation is not easy. For example, there is a lack of electricity, preventing the expansion of the server room. Moreover, the existing server room is not operating correctly. Students even refer to the server room as 'Aoji coal mine,' as sufficient cooling equipment like air conditioning cannot be installed. Therefore, the AI research process is often slow, or results may sometimes not come out at all. Here, the purchase of expensive graphics processing unit (GPU) equipment is challenging, and even if purchased, there are no funds to operate it."
─What do you think are the lessons to learn from China and the U.S.?
"We should benchmark China’s talent development while actively adopting the U.S. industry-academia cooperation model. It means we must liberate ourselves from the shackles restricting industry-academia cooperation. For the country to develop, corporations need to grow, and for corporations to grow, a solid foundation in basic science is needed, creating an organic flow. Our country needs to create an efficient structure from a limited talent pool, but there is a frustrating, high-minded culture present."
"At Stanford University, faculty and students can freely engage in research activities across corporate and academic boundaries without being tied to affiliations, and many of these achievements bear fruit as corporate successes. In contrast, in our country, when professors want to start a startup or hold corporate research positions, there are numerous obstacles such as reviews and regulations, making it difficult even to think about it."
─I heard that there is a serious lack of talent in corporations and institutions.
"I believe the biggest problem currently is that outstanding students are avoiding engineering fields. Soon, Korean talents will start seeking jobs in China. Corporations like Huawei have made massive investments in talent development, with the 139 project being a representative example, offering a shocking deal of hiring one person with a three-year contract at a salary nine times higher. This approach is not easy in South Korea."
"As I hold a professorship, I feel this situation more acutely. Comparing lecture notes from ten years ago to now, it has actually become much easier. To use a sports analogy, there are still excellent talents, but the pool of players is becoming increasingly thin. I see a lack of a compensation system that would attract those talents to Korean corporations. This is also true for semiconductors, which are one of the key industries for the country. Now, semiconductors and AI are converging to the point where it is difficult to discuss them separately."
"Utilizing outstanding entrepreneurs in the public sector, including the government, to devise significant plans is also crucial. Overall, our country has many excellent technicians and leaders. However, too many constraints exist to utilize these individuals nationally, which makes them hesitant to act. To harness exceptional technological leadership for the country, a conducive environment must be created."