As a latecomer in the field of artificial intelligence (AI), South Korea urgently needs to adopt a strategy focused on specialization in select areas to secure competitiveness in the field, according to former Minister of the Office for Government Policy Coordination Koo Yun-cheol.

“We must train AI using data from industries in which Korean companies already have strengths and link that to physical devices such as robots to drive innovation,” Koo said in a recent interview with Maeil Business Newspaper. “Moving forward, our industrial development policies and talent cultivation efforts should also focus on promoting this form of ‘physical AI.’”

Koo, a seasoned policy expert who began his public service career in the Ministry of Finance after passing the 32nd civil service exam in 1989, has held roles such as 2nd Vice Minister of Economy and Finance and Director of the Situation Room at Cheong Wa Dae (the Blue House).

After retiring from public service in 2022, he became a visiting professor in economics at Seoul National University and recently published a book titled AI Korea, which compiles insights from his career.

“Although AI technology is advancing at the speed of light, most people still have only a superficial understanding of it,” Koo noted. “It is fine for only a small number of specialists to handle AI’s technical aspects, but it is urgent to enable ordinary citizens to use AI effectively.”

“When the public can use AI as naturally as they use the Korean language, only then can we expect breakthroughs in innovation and AI-based entrepreneurship,” he added.

He strongly advocated for the “AI+X” strategy, which applies AI across all fields.

This approach involves integrating AI into not only public institutions and companies but also everyday objects to drive work innovation.

Koo particularly emphasized the need to digitize the tacit knowledge of top field experts - knowledge acquired through experience and learning that is difficult to express in language.

“Welding in shipbuilding is a prime example,” he explained. “This area faces significant challenges due to a shortage of skilled workers, and conventional robotic technology has fallen short. If AI-powered robots could learn from experts’ know-how, the impact would be enormous.”

Realizing AI+X, however, requires immediate action from the government, including securing graphic processing units (GPUs), standardizing data, and fostering data science experts.

“GPUs serve as the ‘brains’ that accelerate computation during AI model training and inference. We must secure them even at great cost,” he stressed. “Separately, we must also rapidly standardize industrial data for AI learning.”

Koo also stressed the importance of commercializing new ideas that emerge from daily AI use.

He proposed that both the government and private sector hold AI-related business competitions and provide startup funding for promising ideas while also calling for a structured national AI education system, beginning with early childhood and continuing through primary, secondary, and higher education. He recommended launching online-based “AI education centers” to quickly adapt to ever-changing technologies, rather than establishing new AI departments or graduate schools in universities.

In-person learning needs could be addressed through regional AI technology education centers.

“Teaching people to use AI easily is just the beginning,” he said. “To fully harness AI’s potential, individuals need to internalize its basic operating principles, which will enable the development of personalized learning algorithms and significantly increase AI’s utility.”

Given AI’s massive energy consumption, Koo also emphasized the importance of securing stable power sources and developing advanced materials to reduce energy use. He advocated not only for expanding renewable energy despite its current capacity limitations but for accelerating the commercialization of small modular reactors (SMRs), which offset traditional nuclear energy’s shortcomings, as well. He also urged investment in materials such as silicon carbide (SiC) power semiconductors that can operate under high temperatures and pressure while consuming less electricity.

To prevent misuse of AI, Koo called for the establishment of governance systems. He warned that physical AI, when embedded in robots, cars, and drones, could pose serious security threats if used with malicious intent.

To address this, he emphasized the need for ethics education for humans using AI as well as the introduction of ethical guidelines for AI itself, including the core principle that “AI must not harm humans.”

Koo revealed that he proposed establishing comprehensive guidelines for AI use at the United Nations level at the annual meeting of the Council of Presidents of the United Nations General Assembly (UNCPGA) held in Seoul in 2024.

“We must discuss everything from AI registration and ownership transfer to income, copyrights, and taxation,” he said.

He also suggested that Korea boost its global influence by hosting an UN-affiliated agency akin to the International Atomic Energy Agency (IAEA), tentatively named the International AI Agency (IAIA), among other measures.

[ⓒ 매일경제 & mk.co.kr, 무단 전재, 재배포 및 AI학습 이용 금지]