AI policy and the mismatch between intent and outcomes

Home > Opinion > Columns

AI policy and the mismatch between intent and outcomes

Published: 26 Nov. 2025, 00:01

Audio report: written by reporters, read by AI

Yeom Jae-ho

The author is the president of Taejae University.

Prime Minister Kim Min-seok’s comment at the 2025 JoongAng Forum that “falling behind by one day in the age of artificial intelligence means falling behind by a generation” was a timely message. Countries around the world are investing competitively in AI. But good intentions alone do not guarantee that the expected results will appear.

President Lee Jae Myung speaks at the launch ceremony of the National AI Strategy Committee at Seoul Square in central Seoul on Sept. 8. [YONHAP]

The United States plans to invest $500 billion (736.5 trillion won) in the large-scale AI Stargate project through 2029. The U.S. will build 20 hyperscale AI data centers and pursue an AX strategy that applies AI across industry. It is also moving ahead with market liberalization and regulatory easing for AI development, with the goal of creating 100,000 jobs.

Europe, which had been cautious on AI development, is also changing. In October, the European Commission announced a 340 trillion won investment plan to introduce AI in industry and public administration. France aims to become Europe’s AI powerhouse by investing 109 billion euros ($125.6 billion) in AI infrastructure and by advancing the three-stage strategy introduced in the 2018 Villani Report to apply AI to research, development and industry. German Chancellor Friedrich Merz has recently criticized the fact that AI development is driven mainly by the United States and China, stressing the need for Europe to secure technological sovereignty.

Prime Minister Kim Min-seok delivers congratulatory remarks at the 2025 JoongAng forum at Lotte Hotel in Jung District, central Seoul, on Nov. 19. [KIM KYUNG-ROK]

Combined central and local government R&D budgets in China for 2025 amount to about 800 trillion won. Spending will concentrate on AI, biotechnology, quantum technology and 6G. Alibaba, a private firm, also announced that it will invest 75 trillion won over three years in cloud and AI infrastructure. More than 80 Chinese LLM models — including DeepSeek and Alibaba’s QwQ — are now competing with each other.

The Korean government allocated 10.1 trillion won for AI in next year’s budget. Of that, 2.6 trillion won is for AI adoption in the public sector, and 7.5 trillion won is for cultivating talent. However, this amounts to only 1.4 percent of the country's total 728 trillion won budget. The U.S. alone plans to devote 720 trillion won — 7.7 percent of its federal budget — to AI next year. Similarly, the 800 trillion won China has designated for R&D comes to 15.4 percent of its combined central and provincial budgets.

I had the opportunity in early November at the Beijing Forum to present on AI-based future education with Peking University and other institutions. China is already transforming its educational paradigm through AI, which was striking. In middle and high school mathematics, AI analyzes key content and provides personalized instruction in place of teachers. Edtech companies display dashboards that show students’ learning activity and assessments through graphs and images generated by AI. They have even developed systems that analyze teachers’ lecture voices and present them in graphical form, allowing improvements in pitch and pronunciation.

AI competitiveness is said to depend on three elements: computing centers, data and talent. But what matters even more is a culture willing to adopt these systems, because demand creation is as important as building computing centers. No matter how good a technology is, it will fade if a market does not form to generate demand. Industrial technology policy also emphasizes that demand-pull is more important than technology-push.

The global edtech market is projected to grow at an average annual rate of 15.9 percent from 2025 to 2029, reaching 722 trillion won. Yet Korea’s digital AI textbooks — developed through the dedicated efforts of domestic edtech firms — were not adopted as official textbooks after the change of administration and were reclassified as teaching materials. Even then, the national adoption rate reached only 32.4 percent, ranging from 98.1 percent in Daegu to 24.8 percent in Seoul. Edtech developers say they have suffered losses of several hundred billion won.

School children look at a demonstration for the Booster T1 robot at an event by Booster Robotics in Beijing on Oct. 24. [AP/YONHAP]

Although insufficient prior communication with teachers has been cited, teachers may worry that their roles will shrink if AI textbooks deliver knowledge. In subjects like English and mathematics, the delivery of formal knowledge can now be left to AI textbooks, allowing teachers to focus on fostering broad thinking and curiosity.

Microsoft laid off 15,000 employees last year, most of them software developers or support staff. Coding and programming have become tasks AI performs more effectively. This raises concerns that AI graduate schools funded by the Ministry of Education may be producing software developers rather than the kinds of creative AI specialists the era demands. Just as architects matter more than bricklayers when building a home, creative AI practitioners are more critical than programmers.

To accelerate AI development, Korea must also make active use of diverse data now blocked by privacy and regulatory barriers — including hospital medical data, court precedent data and corporate manufacturing data. Policymakers often speak of “signaling right while turning left.” While the government says it wants to foster AI, regulations in fact appear to be tightening. The mismatch between policy intentions and actual effects needs careful review.

This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.