The reason South Korea ranks second in digital competitiveness globally (according to the World Digital Competitiveness Ranking of the Swiss Institute for Management Development (IMD)) is due to the government’s efforts in digital transformation (DX), such as the establishment of high-speed internet infrastructure 20 years ago. The next step is artificial intelligence transformation (AX), but we have not laid the infrastructure for data centers and algorithm development. We have no choice but to pay the price.
Go Tae-bong, Deputy Minister of iM Securities, noted in a recent interview, “The only way to solve various problems, such as the recovery of manufacturing and the establishment of technological competitiveness, is through productivity recovery, which centers on Sovereign AI and Physical AI. We must approach AI technology pursuit with the mindset of running a marathon at the speed of a 100-meter dash.” He explained that rather than trying to excel in everything that constitutes AI competitiveness, there must be industrial policies that focus on nurturing what we can do well, allowing South Korea to secure at least third place in the global AI competition, following the United States and China.
Throughout the interview, the Deputy Minister raised his voice, stating, “We do not have much time left,” “Choosing and concentrating is the way to survival,” and “AX competitiveness must be secured as a top priority regardless of which government comes into power.” He emphasized that for the troubled South Korean economy to survive again, all citizens must unite to secure AX competitiveness. Below is a Q&A.
South Korean economic productivity is rapidly declining. Should we find the causes of the declining competitiveness of the manufacturing sector in this productivity decline?
“South Korea maintained its DX efforts, such as the establishment of high-speed internet infrastructure, even during the International Monetary Fund (IMF) foreign exchange crisis. We believed this was our way to survive. Thanks to these efforts, South Korea ranks second in digital competitiveness globally among countries with populations over 20 million. There were fruits of investment. However, after DX comes AX, and we have not laid the infrastructure for data centers and algorithm development. We have no choice but to pay the price. Fortunately, no other country, apart from the United States and China, has properly established AX infrastructure. The only way to solve various problems, such as manufacturing recovery and technological competitiveness establishment, is through productivity recovery, which centers on Sovereign AI and Physical AI.”
What is the relationship between securing manufacturing competitiveness and improving productivity with AI?
“Physical AI is literally AI technology that has been implemented in real physical environments. When it comes to improving productivity, many mention ‘relaxing the 52-hour workweek’ and ‘flexibilizing the labor market.’ However, once Physical AI is introduced, such discussions become meaningless. Robots or machines equipped with AI do not need to adhere to the 52-hour workweek and do not go on strike. We must accelerate the development of Physical AI while maintaining our manufacturing capabilities.”
What does it mean to maintain manufacturing capabilities?
“Consider the war between Ukraine and Russia. The Western countries, including the United States, have AI technology but lack manufacturing capabilities, which prevents them from producing Physical AIs like AI drones or AI submarines. In contrast, China, which has manufacturing capabilities, is producing unmanned drones and unmanned submarines. It is essential to have manufacturing capabilities to transfer advanced AI technology into the physical world. Fortunately, South Korea still possesses manufacturing competitiveness. In particular, our shipbuilding technology, crucial to naval power, remains at the world’s top level. We must develop AI technology while maintaining manufacturing capabilities to create Physical AI.”
Labor is decreasing due to low birth rates and an aging population. How should we approach the issue of technological competitiveness?
“AI can also be a solution to technological competitiveness issues. There are two ways to enhance AI competitiveness: partnership with leading firms and independent development. However, there are limitations to collaborating with leading firms. For instance, China is the world leader in drone technology, making it difficult to partner with Chinese companies for military drones. Therefore, independent development is necessary, but already the technological gap has widened too much. Ultimately, given our reality, we must pursue both technological cooperation and independent development. The key is that we need to secure our own AI technology, Sovereign AI. Even if we partner with leading firms overseas, having our own technology is essential to respond to any situation. Above all, in areas like defense, where there is a risk of information leakage, the development of Sovereign AI is crucial.”
How should labor market policies change to enhance competitiveness in high-tech industries?
“We cannot win by directly competing with the U.S. capital and China’s 1.4 billion workforce. Therefore, we need to focus on what we do well. Labor market policies must also reflect this. The labor market should provide the time to concentrate through flexibility. I believe a focused work system is needed where we work intensively when we work, and rest sufficiently when we are not working.”
What changes are necessary in education policy?
“Education policy, in other words, is about securing talent. We need to look at programs like China’s ‘Thousand Talents Program,’ initiated by the government since 2008 to attract overseas scientific and technological experts. The recent ‘Gold Card’ program by the Trump administration, which sells U.S. residency for $5 million (about 700 million won), is similar. Some critics question, ‘How can you sell residency for money?’ but it can be utilized effectively by U.S. corporations to attract talent. If interpreted as a willingness to bring in talent from around the world at a cost of 700 million won, it could be viewed differently. Fortunately, we also have many talents who have studied advanced sciences abroad. We need a bold policy that can attract them.”
A new government will be established in June. What kind of industrial policy should be pursued?
“AX competitiveness must be secured as a top priority, regardless of which government enters power. We must pursue AI technology with the mindset of running a marathon at the speed of a 100-meter dash. We do not have much time left. If we fail this time, there will be no opportunity. We must secure AI competitiveness before the window of opportunity closes. The United States and China are already far ahead in the AI competition. The good news is that with focus and concentration, securing at least third place is feasible. Apart from the U.S. and China, no other countries have truly secured their AI competitiveness. Therefore, we should implement a strategy to chase the U.S. and China while surpassing Germany and Japan.”
[keyword]
Sovereign AI refers to an independent AI that is not subordinate to any specific country. In the process of developing and using AI models, vast amounts of data concentrate on AI. However, there are concerns that this data, if leaked, could pose a critical threat to national security, leading to arguments for developing Sovereign AI, even if it involves somewhat lower technology, to apply it in areas closely linked to national defense and the public sector.
Plus Point
“The next step after Generative AI is Physical AI... This year will be the beginning.”
“The next stage of Generative AI will be Physical AI, which moves physical entities like robots and cars.” Jensen Huang, CEO of NVIDIA, said this during a keynote speech at the Consumer Electronics Show (CES 2025), the world’s largest IT and electronics expo, held in Las Vegas, U.S., in January 2025. He defined Physical AI, which changes life in directly visible ways like robots and autonomous driving, emphasizing it as a ‘next wave.’ He also projected that this year will be the inaugural year for Physical AI.
Physical AI refers to technology in which AI, which used to exist only inside computers or smartphones, is now integrated into machines like robots, allowing interaction and movement in the real world. Previously, existing Generative AI would convey knowledge primarily through conversation with humans. In contrast, Physical AI can implement AI in real physical forms. The representative technology of Physical AI is humanoids (human-shaped robots). Humanoids are referred to as Physical AI for this reason, embodying a cutting-edge fusion of AI with advanced machine technologies such as sensors, precise control, and batteries.
Through bodily intelligence that implements precise physical strength at the human level, Physical AI can perform activities in the real world on behalf of humans. It is argued that in order for Generative AI to evolve into Physical AI, it must go through an intermediate phase called AI agents. AI agents refer to algorithms that autonomously plan and execute after understanding the problem. While existing Generative AI operates based on input training data, AI agents represent an intermediate technology combining the software aspect of AI and the hardware of humanoids. The key question is whether AI will make decisions directly on behalf of humans. In contrast, Physical AI fundamentally makes decisions and executes actions by itself.