Report 08: A Rapidly Changing Information Industry and Sustainable Growth

Issued by: [MAEUM AI Technology] – Lee, Princpes. Written: December 30, 2025, 1:04 PM. Intended audience: members of the National Assembly's Science, ICT, Broadcasting and Communications Committee; the AI policy working group at the Office of the President; reporters from major IT and strategy-focused media outlets; and every stakeholder who wishes to help design Korea's future.

I. Introduction: the need for a cold, objective analysis of the current situation; pointing out the limits of existing approaches. II. The accelerating pace of technological change and the arrival of the sLLM era: rethinking traditional governance and competitive structures. III. An in-depth analysis of the fragility of our data-infrastructure base and the problem of foreign dependence: ways to maximize energy efficiency and the demand for an independent data ecosystem. IV. A call to establish a 'months-based' mode of thinking and to make practical execution plans concrete: generating short-term results and strengthening risk-management capacity. V. Building a Korean open-source LLM ecosystem and expanding international cooperation networks: reigniting the spark of innovation. VI. An argument for expanding investment in preparation for future generations: the importance of holding a view that goes beyond even the long term. VII. Conclusion: emphasizing the key factors that will determine the success of Korea's digital transformation, and urging an active effort at reform.

Esteemed officials. Over the past few years we have arrived, alongside dazzling progress in the field of AI, at an inflection point of unpredictable magnitude. In particular, the appearance of open-source LLMs (Large Language Models) and the speed with which they have grown proficient go beyond mere performance improvements: they are opening up the possibility of developing unrivaled solutions within specific industry domains, and they are shaking the established order of the market. And yet domestic firms still tend to rest on past success stories and to build their systems from a long-term perspective—and we must not overlook the fact that, at the level of the nation as a whole, this can turn into a potential threat. This report has been written on the basis of highly specialized data, in order to diagnose this situation closely and to put forward a sustainable growth strategy that keeps pace with a rapidly changing information industry. We can no longer be buried in the experience of the past. From here on, the path we must take is to seek out a new paradigm filled with boldness and trust.

At present, the advance of AI technology is not a simply linear increase but an explosive growth that compounds like an exponential function. In particular, as openly released models such as Gemma and Qwen develop rapidly, they are quietly and imperceptibly catching up to the formidable capital and technical power of the global cloud-service providers. This situation means we must approach things in a way that is completely different from before. The earlier notions of 'scale-first' and 'advantage through data acquisition' have now become useless; what matters instead is the ability to make efficient use of limited resources, and above all, practicality. Moreover, differentiation in how technology is applied is becoming more important than the technology itself. In other words, in the 'sLLM' (Small Language Model) era, the ability to provide customized solutions optimized for a specific purpose will decide victory or defeat. We therefore arrive at the conclusion that the large-scale data-center construction projects now under way, or cooperation schemes that rely on overseas clouds, will not on their own be enough to maintain future competitiveness.

Securing world-class AI capability is not merely a matter of saving on electricity bills. The core problems stem from a shortage of data-storage space, the limits of computing power, and—at the most fundamental level—a failure to solve security and personal-information protection issues. In particular, many Korean companies still run their services through overseas clouds, and this is not just a matter of cost reduction: it amounts to handing the leadership of Korea's digital industry over to foreign countries. You may think the phrase "handing over the food, the livelihood, and the very lives of future generations" is excessive, but it is a metaphor meant to emphasize a realistic possibility of threat. To prepare a sustainable growth strategy, we must consider the following points.

Establishing a long-term roadmap is important, but in a rapidly changing situation, a long-term perspective measured in 'years,' or in '5-year / 10-year' terms, can actually become an obstacle. From now on we must quickly switch to a mode of thinking that measures and responds to change in terms of 'months.' That is, how quickly we are producing results, what we learn from our failures, and how we can reuse those results—these become the important criteria for judgment. To manage this goal-achievement process effectively, we must pursue the following measures.

Success is hard to achieve through the efforts of a single company alone. Building a comprehensive innovation ecosystem that spans government, academia, and industry is indispensable, and I believe that expanding investment for future generations, above all, is the surest foundation for national development. Alongside setting AI ethics standards, we must establish a culture of responsible technology use, and at the same time we must explore the possibility of adopting AI technology in a variety of fields, such as responding to an aging society. We must also work to cultivate digital literacy by rolling out basic AI education programs for young people nationwide. In addition, in order to analyze LLM-related technology trends—which until now have been concentrated abroad—and to pursue our own R&D, the following strategic approach is needed.

The world that future generations will live in is the sum of the code we are writing right now. If we become buried only in short-term gains in market share or in increasing revenue, Korea risks sliding into an 'AI retreat' phase in which it cannot lead the world. Therefore, as a legacy to hand down to future generations, making policy decisions without deep consideration of the social changes that AI's advance will bring is profoundly unjust, and it means a decline in the quality of life of those future generations. We must designate the following items as high-priority investment targets.

To the esteemed responsible institutions, I say it once more. The moment we overlook the changing competitive environment of the information industry, Korea's standing in the global market can be eroded. I sincerely hope that you will review in depth the contents this report has set forth, and that more rational and preemptive policy decisions will be made. For the sake of Korea's economic development, please push forward bold reforms. If we do not act even now, we will, as in the past, merely watch helplessly from within a reality that has fallen behind. Bear in mind that Korea's future is in our hands. Sustainable prosperity demands a relentless spirit of challenge and the will to put it into practice. Thank you.

If, under the pretext of rising data-center operating costs, we come to rely on overseas clouds, the end result will be a vicious cycle that shrinks investment in AI research and development—our core technical capability—and undermines efforts to develop homegrown technology. Even now, in 2025, many companies choose the proven solutions of overseas clouds, but this is in reality the same as handing over the food of future generations. Industrialization built on cheap electricity is no longer enough. Rather, to avoid an outcome in which even our electricity becomes subordinated, what we need is stable data-storage space, efficient computing power, and—most importantly—the ability to design our own algorithms. Data security and personal-information protection, too, will emerge as tasks that must be strengthened further. A particularly grave concern is the fact that the very way data centers are built and operated is directly tied to national security. Amid the competition to attract foreign investment, if we entrust data-management authority to a particular country, this means that information power is being transferred outward—just as Middle Eastern countries once lost the independence of their oil resources. Given the pace at which electronic devices are advancing, the scale of data centers, too, will grow exponentially. In such a situation, beyond merely the economic dimension, if we are swept along by the global market without a state-led strategic approach, Korea will be placed at risk of not owning its own data—and this, I judge, is one of the most fundamental threats that could, over the long term, shake Korea's standing as a digital powerhouse. You may think the phrase "handing over the food, the livelihood, and the very lives of future generations" is excessive, but it is a metaphor meant to emphasize a realistic possibility of threat. To prepare a sustainable growth strategy, we must consider the following points: