[AI Rule Maker]④ Kwangjae Lee: "Expanding AI Integration in National Assembly Budget Review... Scaling Up Investment to 50 Billion Won"

Interview with Kwangjae Lee, Chairman of the National Assembly Special Committee on Budget and Accounts

Tracking Carry-Over and Unused Budgets Through an "AI National Assembly"

Sorting Out Projects that Are "AI in Name Only"

GPUs Are the Strategic Weapons of the AI Era – They Must Be Allocated at the National Level

Kwangjae Lee, Chairman of the National Assembly Special Committee on Budget and Accounts, has announced that artificial intelligence (AI) will be actively introduced in the review of the national budget bill and that related investments will be scaled up to 50 billion won. The plan is to analyze the validity and execution performance of government projects using AI and to improve the customary practice of repetitive, similar questions during the budget review process.

During his tenure as Secretary General of the National Assembly, Chairman Lee spearheaded the establishment of an "AI National Assembly" that supports data-driven legislative activities and pursued investments of over 50 billion won. However, he diagnoses that the actual project size was later reduced to the 10 billion won range, resulting in the system falling short of the originally envisioned standard. Now that he has resumed the chairmanship of the Budget Committee this year, he plans to restore the "AI National Assembly" budget to the 50 billion won level in next year’s budget bill review process.

Kwangjae Lee, Chairman of the National Assembly Special Committee on Budget and Accounts, is being interviewed by The Asia Business Daily at the National Assembly. August 18, 2026. Photo by Hyunmin Kim

In an interview with The Asia Business Daily, Chairman Lee stated, "If we analyze decades of accumulated national data and National Assembly transcripts with AI, we can reduce repetitive and redundant questions," adding, "We need to boldly increase the related budget next year." He further emphasized, "For reviewing a national budget of 800 trillion won, an investment of 50 billion won in an AI system is by no means a large sum."

As for the anticipated effects of AI adoption, he explained, "Pre-feasibility studies and project analysis and evaluation will become faster, and we’ll be able to quantify agendas that are directly linked to people’s lives." He said he plans to prioritize analyzing the effectiveness and bottlenecks of budget allocations in areas such as delayed housing supply, youth employment, and low birthrate and aging population projects.

He proposed that AI-related budgets should be managed in an integrated way at the national level rather than being dispersed across ministries. Chairman Lee said, "It is essential to establish a unified management system to efficiently allocate GPUs, which are like strategic weapons in the AI era," and added, "We will exclude projects that are 'AI in name only' and focus on reviewing budgets for projects that can deliver practical services and added value in people’s daily lives."

The following are questions and answers from the interview with Chairman Lee:

- In the context of the US-China competition for AI leadership, where should Korea’s competitiveness lie?

▲ AI is not just about the emergence of a technology, but a civilizational transformation. It is imperative that we recognize success or failure here will determine the fate not just of individuals, but of the nation. Ultimately, Korea must become the country that uses and leverages AI better than anyone else, from its citizens to its companies. Just as we can no longer imagine life without computers or smartphones, soon there will be a world where we cannot live without AI.

- What do you see as the biggest bottlenecks in Korea’s current AI ecosystem?

▲ All three: data, computing infrastructure, and energy, must be resolved. Data only becomes meaningful when it is converted to an Application Programming Interface (API). Personal information should be protected, but APIs should be boldly made open, and the network separation issue needs to be addressed. For data centers, we need to expand undersea cables and low-orbit satellite networks. From an energy perspective, not only renewable energy but also small modular reactors (SMR) must be pursued vigorously.

- Do you mean that regulations on personal information should be relaxed to promote data utilization?

▲ That’s correct. For example, in Korea, the National Health Insurance Service and Health Insurance Review & Assessment Service hold vast amounts of medical data, but data from hospitals and individuals is not integrated. Each hospital uses different software, creating incompatibility. If we use this data properly, it can even facilitate cancer diagnoses based on medical imaging records.

Currently, Korea neither protects personal data effectively nor allows its use, creating a country that blocks data under the pretext of protection. The US strictly protects personal information but also uses data. Europe’s approach, which has strengthened regulation under the banner of data protection, has caused them to fall behind in platform development. We need to more actively embrace the American model.

- Do you believe compensation is needed for citizens who provide data?

▲ There must be proper incentives for sharing personal data. Just as there is a life insurance company that reduces premiums for people who walk more, those who provide data should be rewarded. Currently, only those who take the data benefit, with no rewards for those who create it. Whether the state or the private sector, rewards in the form of insurance premium discounts or points must be given to citizens who provide data so everyone can benefit.

- Where should the government's AI-related budget be prioritized?

▲ First, we need a system to effectively distribute GPUs. Globally, it is said that Nobel laureates typically use about 500 GPUs for their research. The issue cannot be solved by handing out a few dozen here or there. GPUs are a kind of strategic weapon. Instead of allocating them unit by unit, the country should strategically manage them and operate shared research facilities, allowing those who need them to use them as needed.

Additionally, Korea is a country strong in application technologies. By integrating AI into our world-class manufacturing, we can become a manufacturing powerhouse. AI in education should enable every citizen to have a personalized tutor. In healthcare, each individual could have a dedicated physician; in culture, we could create global platforms where numerous writers and creators can earn income. Currently, we are paying 'digital rent' for apps, cloud, and AI services. We must secure platforms of our own.

- What standards should be used to review AI-related budgets, currently scattered across ministries and projects?

▲ First, we must manage whether GPUs are being used efficiently and create systems to predict their outcomes. Second, we must closely check whether there are projects that are 'AI in name only' where the term is used indiscriminately. Third, we must focus on whether the projects can provide useful real-life services and create added value. For AI as well, rather than just the technology itself, services that are actually useful for the public must emerge. Only then will new sectors, companies, and jobs be created.

- What is required to foster AI talent and ensure continuity in R&D investment?

▲ Institutions and laws should ensure that policies are not disrupted every time there is a change in government. A national system is necessary that selects challenging R&D projects, even if the probability of success is low. The visa system for recruiting top foreign talent should be radically reformed. Initially, researchers should be allowed to start with a small budget, and if potential is demonstrated, the government should procure on a pilot basis and then scale up in the market as needed.

- There are concerns that AI investment results may be concentrated among a few large conglomerates.

▲ In the AI era, there will emerge businesses that are like multinational corporations or huge empires, and wealth is likely to be concentrated among a few companies. The AI era is comparable to a kind of digital feudalism. Just as irrigation channels in the agrarian age or electricity and water supply in the industrial era were universal infrastructure, the state must provide digital infrastructure enabling everyone to use AI easily.

The shareholder return policies by Samsung Electronics and SK hynix are necessary, but I believe it is too early to celebrate. Given the fact that state infrastructure such as electricity and water has been provided, companies must increase youth employment, make supply prices more realistic for partners, and conduct joint R&D. Companies cannot last unless there is a belief that the entire public grows together, not just single corporations.

Kwangjae Lee, Chairman of the National Assembly Budget and Accounts Special Committee, is giving an interview to The Asia Business Daily at the National Assembly. August 18, 2026. Photo by Hyunmin Kim

This content was produced with the assistance of AI translation services.