Growth requires strength. Strength that will not waver, strength to reach even higher. "Take Care of Culture," a knowledge news program that makes you smarter the more you listen, empowers you to grow as a cultured person alongside top SPOTV contributors.
Contributor Bae Kyung-hoon
Deputy Prime Minister and Minister of Science and ICT
Former Head of LG AI Research
Current Vice Chairman of the National AI Strategy Committee
These days, I am largely contemplating two key questions: First, how can South Korea produce world-class AI technologies and services so that the entire world utilizes South Korea's AI? Second, to reduce the growing AI digital divide, how can we effectively roll out initiatives such as providing AI education and giving the general public easy access to AI, thereby narrowing this gap?
Can South Korea Truly Become an AI Powerhouse?
Q. With America's Nvidia seeming to take over the world and China pouring all its national might into AI, South Korea appears to lack both capital and infrastructure. Can we hold our ground?
South Korea ranks first globally in industrial robot adoption. In addition, when looking at the per capita ratio worldwide, South Korea has the highest generative AI usage rate.
Sam Altman | CEO of OpenAI (July 24, 2026)
The South Korean economy and its people have embraced AI remarkably fast. The transformation unfolding in South Korea is truly astounding.
South Korea has an immense interest in the AI era. It is true that investments related to artificial intelligence demand staggering costs. In the United States, big tech companies have led AI investments for a long time, and China has also invested heavily in AI over the past decade to keep pace in the U.S.-China tech hegemony race. It is true that South Korea, despite having less capital, has shown tremendous interest in AI. What comes to mind when you think of South Korea? A powerhouse in semiconductors and manufacturing. Because the government is investing heavily, businesses have a strong commitment to AI transformation, and solid foundations are already established, I believe we can build AI effectively on top of that base.
I do not believe having a large population or a massive difference in capital is the deciding factor. LLMs are the product of training based on massive datasets. Researchers worldwide use Nvidia's GPUs for AI training, but exports to China are restricted. China's AI models are achieving their current results even within poor AI infrastructure environments. While investment is certainly correlated to some extent, the key to making AI smarter lies in data. Therefore, the focus is on how to gather high-quality data. Most of the knowledge required for the LLM era has already been learned. I believe we have reached the point where we must create points of differentiation. In that sense, I see South Korea's manufacturing-based data and capabilities playing a critical role in the AI era.
Q. Will South Korea be competitive if it develops models independently?
I will break this down into AI model competitiveness and AI service competitiveness. What could happen if we fail to secure independent AI model competitiveness? Recently, the U.S. government restricted other countries from using Anthropic's top-tier AI. If researchers working with the latest AI model suddenly face a cutoff in its supply, continuing that research becomes extremely difficult. Relying on older versions of AI models inevitably leaves one lagging behind in competition. Such situations can occur at any time. Reports have also emerged that China plans to restrict overseas supply. Since cutting-edge AI models are the result of massive investments, some may impose restrictions for profitability reasons, or they may choose not to make them public as core assets for national security and defense. If that happens, the gap between nations possessing the latest AI models and those that do not will widen dramatically. That is why we must possess the competitive capability to build AI models of a certain standard, and South Korea is working to secure competitiveness through its proprietary AI foundation model* initiative.
*Foundation model: An artificial intelligence (AI) model trained on vast amounts of data that can be applied to a wide range of tasks
In fact, the U.S.-China rivalry for tech hegemony can create great opportunities for South Korea. AI technology seems somewhat different from physical products. Because it can be described as a trained aggregate of massive knowledge, relying solely on someone else for AI can be risky. For this reason, South Korea is investing heavily in AI as a third alternative, and indeed many countries believe South Korea is the only nation outside the U.S. and China capable of developing its own sovereign AI. OpenAI CEO Sam Altman once said, "There is no AI revolution without South Korea."
Sam Altman | CEO of OpenAI (July 25, 2026)
Without South Korea, today's incredible AI revolution would not have been possible. It is truly remarkable, and I am confident there will be even more changes ahead.
South Koreans have the highest AI usage rate in the world. As such, South Korea is leading the revolution of this AI era. Nvidia's Jensen Huang also noted, "Right now is South Korea's golden age of AI."
Jensen Huang | CEO of Nvidia (July 25, 2026)
This is South Korea's golden age. As you know, South Korea's semiconductor business is booming, and its industries are thriving.
I believe this reflects South Korea's strong potential to serve as an AI production base, backed by strong competitiveness in memory semiconductors, AI chips, and AI models. Moreover, South Korea is not trying to monopolize AI, but is actively engaging in discussions to cooperate with fellow middle powers
. Because many nations view South Korea as a country that possesses the full AI stack
and can collaborate to build an AI ecosystem together, they are seeking cooperation discussions.
*Middle power: A nation in international politics that possesses intermediate state power and influence between major powers and small states
*AI full-stack: Encompassing the entire spectrum of related industries, from hardware infrastructure including GPUs to software, cloud services, and AI solutions
Physical AI, the 'Ultimate Stage' of AI: Why South Korea Aims for the World's Best
Q. You cited Physical AI as the ultimate destination in the AI race. Why do you consider Physical AI the next battleground for the AI industry, and what is South Korea's greatest competitive advantage in that process?
I believe Physical AI represents the ultimate stage of artificial intelligence. Conventional AI existed solely inside computers, and people accessed AI services through smart devices. Now, that AI is stepping out into the real world to operate. Beyond humanoid robots, Physical AI can be applied across diverse fields, including autonomous vehicles, drones, factory robots, agricultural robots, and caregiving robots for the elderly. If generative AI signaled the start of the AI revolution, Physical AI is what brings it to full fruition. It will create massive industrial impact, and there is just as much preparation required.
The reason I believe South Korea can excel in Physical AI comes down to data. LLM technology, represented by generative AI, learned by gathering massive amounts of human-accumulated data. Generative AI was built by incorporating the books, academic papers, and know-how we created. It reached today's results through more than a decade of collecting foundational data, defining data attributes, and labeling them.
However, Physical AI requires action data so robots can accomplish missions, and such data is scarce. While robots can mimic human actions by learning from video clips, that is still at the stage of performing specific simple missions. Because South Korea spans full industrial sectors needed for Physical AI—including automobiles, semiconductors, shipbuilding, and bio—it can lay the groundwork for Physical AI based on that data.
To make Physical AI fully operational, such industrial data is drastically insufficient. Without adequate data, Physical AI cannot operate. Compared to the colossal amount of data trained to build generative AI, we don't even have one-millionth or one-ten-millionth of that data. Skilled experts in each field need to collaborate with AI or robots to generate this data.
Take a tennis ball and a table tennis ball. Their weights differ. Their sizes differ. If you throw them, which one travels farther? The tennis ball. Why? It weighs more and has a higher coefficient of friction with the hand, generating a stronger reaction force that carries it farther. Data must be created taking all these physical laws into account. We currently live in the physical world by accounting for various laws of gravity and reaction forces. AI must also gain a physical understanding of the real world to perform actions or missions like a human.
Understanding the physical real world is not the end of it; the AI must actually move. It has to learn from someone. For example, a robot taught table tennis by me would lose every single match. (Laughs) But if it learns from behavioral data obtained from a professional table tennis player, it will quickly become a robot capable of playing table tennis. Those things are what matter.
To excel at Physical AI, we must effectively collect relevant Physical AI data and establish a sound framework to secure it. Because South Korea already possesses the foundation, we can advance swiftly. In the field of Physical AI, South Korea must strive to become the world's best.
Q. Having master craftsmen in manufacturing who are comparable to elite baseball pitchers is itself a global competitive edge. That allows us to acquire data quickly.
We will be able to build systems to gather data across industries, and we could also have workers wear motion suits to operate robots and acquire data for specific industrial missions. I view this as an elementary stage of Physical AI. For Physical AI to be truly complete, robots must not only accomplish tasks assigned by humans, but also understand situations on their own, make judgments, and reach decisions. To advance to that level, experts must participate in creating data for Physical AI as well. Securing high-difficulty, high-quality data is the core of Physical AI competitiveness.
What I am envisioning right now is transforming entire factories into Physical AI. When people think of humanoid robots, they usually only imagine carrying objects and performing tasks. When an entire factory becomes Physical AI, AI handles production design independently, places component orders on its own, and assigns respective missions to each robot inside the plant. AI makes optimal decisions through comprehensive evaluations, and conducts one-stop quality inspections on products and overall facility checks. South Korea will create factories that are themselves AI. We are currently conducting pilot verification projects, and we are working to export these factory AIs as complete packages. This is real.
Making Money with AI You Created? The New Future Reshaped by the 'Token Economy'
Q. For us to become the global No. 1 in Physical AI, the term "token economy" has been coming up often lately. Does a token economy mean selling data created by manufacturing masters?
We could attach value to and pay for data created by masters across various industrial sectors, or compensate masters for their know-how when they take part in generating synthetic data for Physical AI. But the concept of the token economy as I see it is that tokens are the fundamental unit that AI understands and produces. You can think of an AI data center essentially as a factory that generates tokens. A token factory*.
*Token factory: A system that combines GPUs, high-speed networks, storage devices, and AI software to produce tokens (the outputs of intelligence) on a massive scale
Operating Physical AI also requires tokens. AI services have a set amount of tokens, and once they are exhausted, you cannot use them further. You can think of all of that as tokens. Currently, everyone is merely consuming tokens. However, if a Physical AI built by me through an AI agent I created performs meaningful activities, and the outputs of my activities receive meaningful feedback or compensation, then while we consume AI, we can also become producers of tokens through AI. I would describe this as Intelligence Token Economics*.
*Intelligence Token Economics: A structure in which 'intelligence' generated by AI is produced and consumed in token units, with its costs and values driving the economy
Talk of an AI bubble in the United States right now centers on whether the returns match the scale of investment. That is why the stock market keeps fluctuating. It happens because a positive virtuous cycle has not yet been established. We invest in memory, AI chips, and GPUs, but the question is whether productive effects matching those investments are materializing. I believe building such an AI ecosystem effectively is ultimately paramount. An AI ecosystem should not be limited to the role of specific corporations that develop AI infrastructure, models, or services. From the user's perspective, if a virtuous cycle is established where users produce content through AI or participate directly in socioeconomic activities to receive rewards or feedback as content producers, a virtuous economic cycle will be created.
Through "AI for Everyone*," which launches this year, our goal goes beyond having every citizen use AI for Everyone; we aim to build a token economy where people receive rewards for data and content generated through AI for Everyone, and when they exhaust their tokens, they can continue using AI services with additional tokens. This concept has not yet been established globally. South Korea is paying keen attention so that it can swiftly introduce and successfully establish this model.
*AI for Everyone
-A domestic AI service available free of charge to all citizens without usage limits
-General-purpose AI chatbots and public AI agents scheduled for release
Q. Given the current internal circumstances in South Korea, is there real potential and competitiveness?
I believe there is. A perspective I have maintained since my time in the corporate sector is that applying AI from inside computers to on-site operations involves substantial trial and error. In fact, most give up at this stage. However, AI models themselves are probabilistic models. They may not score 100 points right from the start. Most AI model services need a process of first creating a 70-to-80-point model and gradually improving it to 80, 90, and 100 points. Sustained investment over time is necessary. I also wrestled with this issue when I was in the corporate sector. From an investor's perspective, they want short-term results, but without going through this process sufficiently, producing a final product is difficult.
Now, as I shape AI policy, I feel full of anticipation. Interest and investment in AI this year have grown compared to last year, and they will expand even further next year. They will be strengthened further. When the government announced the Three Mega Projects, the announced investment amount alone exceeded 4,000 trillion won. At the recent San Francisco AI Summit, investment discussions between big tech firms and leading South Korean AI companies totaled $950 billion alone. Staggering investments and bets are centering on South Korea.
Earlier this year, a reputable institution evaluated that "South Korea distinctly possesses top-three AI capabilities globally," demonstrating that we are also building AI model competitiveness. While model competitiveness is vital, we must now build service competitiveness to create AI that citizens and businesses can actively use, and citizens and businesses have reached the stage where they need to use it. By utilizing it extensively and leaving good feedback, AI for Everyone will not be something provided by the government, but something created together. We can build the platform together, expand its scale, and take on roles not just as consumers, but as producers. Once such success stories are forged in South Korea, they will spread worldwide. What matters now is actual usage.
"In the AI Era, the Gap Widens Further": What Solutions Is South Korea Preparing?
Q. In the AI era, the gap between those who utilize AI effectively and those who do not could widen. How do you view this AI divide?
I view it very seriously. In reality, I believe the gap will continue to widen. The divide between those who use AI and those who do not, as well as between those who use advanced AI and those who use low-level AI, will grow even larger because the resulting outputs differ markedly. We have reached a world where one can no longer think, "I can live without AI." Everyone needs to utilize AI, yet many people still lack touchpoints and opportunities.
I believe AI models and their outputs should essentially become universal public goods worldwide. They should be integrated into daily life like electricity so anyone can use them. However, because we are still at the beginning of the great AI transformation, colossal investments must be recouped. Consequently, the divide between those who supply and excel at using AI, those who use AI passively, and those who have no access to AI at all is bound to widen. If we do not narrow that divide, I believe massive social problems will arise. Therefore, to ensure no South Korean citizen is left behind from AI benefits, we are paying great attention to providing AI that all citizens can use without financial burden, while simultaneously building a virtuous AI token economy.
Everyone uses ChatGPT or Gemini, right? You probably don't use South Korean AI, do you? That is not because South Korean AI models do not exist, but partly because popularized services have not yet emerged. That is why the government and industry are joining forces to launch a service called AI for Everyone starting at the end of the year. Currently, citizens have to search for and apply for community center services or government programs individually. We plan to launch AI for Everyone this December to enable all citizens to access services easily by building public AI agents and combining generative AI services with public AI agent services to create unique differentiators. While there may be shortcomings compared to foreign AI services, our citizens need to use domestic AI services extensively. With widespread usage, data continues to accumulate due to the nature of AI models, enabling continuous advancement. Not only the government but many companies are preparing for this together with great interest.
There are also many programs provided and educated at the government level. For example, through the AI for Everyone Contest or the AI for Everyone Learning Center, we aim to provide all citizens with diverse experiences through AI and allow them to realize they can accomplish many things they had never conceived before through AI. Ensuring all citizens can participate without being marginalized is one of our key goals. While the divide will grow, it is essential that we contemplate ways to narrow it, formulate policies, and strengthen AI literacy education.
Q. When people hear "AI competition," it might feel like an event strictly for developers. What makes this National AI Contest different, and what kind of competition is it?
People feel apprehensive simply because they have never used AI before. We want to provide the experience of making life richer and more valuable through artificial intelligence. The AI for Everyone Contest has already surpassed its target of 2 million people this year, and we will continue providing educational opportunities and running the competition. We will raise the targets next year to build the foundation for people to experience AI.
Go Tae-hyun | Department of Tax and Accounting, Hyupsung University
After using various AI tools, I learned about this year's "AI Rookie" competition for university students and decided to participate because I wanted to develop a finance-related AI.
Lim Chan-young | Department of Computer Software, Korea Bible University
I took on this challenge through this project because I wanted to write a research paper.
We have seniors over 70 publishing books and active as YouTubers based on AI content. Elementary school students are also participating in the AI for Everyone Contest by creating AI outputs. They gain new knowledge and feel a great sense of accomplishment. They say things like, "Can I really create something like this?" Even if you wanted to write a book about your life, it would require a vast amount of time without AI, but with AI you can craft your story and easily create illustrations. Through these experiences, people say, "We are truly living in a whole new world."
What is certain is that the era of great AI transformation will come. The AI revolution has already occurred, and I believe a new world many times, hundreds of times greater than the past ICT revolution, semiconductor revolution, and internet revolution will be forged by the AI revolution. Rather than worrying, "Will my job disappear because of AI?", I hope young workers in their 20s-30s and 30s-50s will first try using AI in areas of interest or challenge. I hope we can collectively explore how to learn with AI, solve problems, and grow in the AI era. The government will deliberate and support alongside you, and we will join hands to run together so South Korea can navigate the era of great AI transformation successfully.
Subscribe to the "Take Care of Culture" YouTube channel to see more content.
※
Copying, redistribution, and unauthorized use in AI training are strictly prohibited.