Kim Eun-mi
The author is a professor of communications at Seoul National University.
Discussions about artificial intelligence are everywhere in Korea. Seminars and lectures on how to use AI are multiplying. Some expert assessments suggest that current AI systems operate at an equivalent of an IQ of 120 to 130, well above average human intelligence. With such powerful tools in hand, the opportunity cost of spending time acquiring and processing information independently has risen sharply.
An official from the Ministry of Science and ICT delivers the opening speech at the “2025 Korea AI Education Festival” at KINTEX Exhibition Center II in Ilsan, Goyang, Gyeonggi, on Nov. 21, 2025. [NEWS1]
Korea’s rapid embrace of AI reflects its longstanding dynamism. Much like the swift expansion of high-speed internet in the late 1990s, there is now a sense of urgency to integrate AI across all sectors. Yet the question remains whether education should follow the same pace. Unlike other fields, mistakes in education cannot easily be undone.
A recent experiment by Anthropic, which developed the AI model Claude, offers a useful insight. Participants were divided into two groups and given coding tasks, with only one group allowed to use AI. As expected, the AI-assisted group completed the tasks more quickly. The more revealing result came in a follow-up experiment. When both groups were later allowed to use AI, the group that had initially worked without assistance outperformed the other.
The explanation is simple. Those who struggled through the first task without technological help developed stronger problem-solving skills. Their independent efforts enhanced their capacity to think and adapt. The result suggests that prolonged reliance on AI can hinder learning, much like trying to learn to ride a bicycle while depending too long on training wheels.
These findings highlight the need for a careful approach to introducing AI in education. Unlike in other sectors, speed alone should not be the priority. Policymakers and educators must strike a balance between encouraging AI proficiency and preserving motivation and intellectual curiosity. True AI literacy includes knowing when not to use AI.
Another concern is that rapid adoption of AI in education could deepen existing inequalities. Korea once led the world in high-speed internet access and there was widespread belief that computer education was essential to prevent students from falling behind. At the same time, the experience revealed unintended consequences. Without proper guidance or digital literacy, some children became overly dependent on online games or developed unhealthy usage patterns.
This history offers a clear lesson. The impact of technology depends less on its adoption than on how it is used. The same principle applies to AI. Simply introducing the technology into classrooms does not guarantee better learning outcomes.
A company representative demonstrates humanoid robot movements at the “2026 Seoul National University AI Summit” at the Global Engineering Education Center of Seoul National University in Seoul on April 2. [YONHAP]
There is also concern about how AI’s instant-answer model may interact with Korea’s exam-driven education system. Students are expected to develop persistence and critical thinking by solving problems and learning from mistakes. Yet if assignments produced with minimal input through AI receive higher evaluations, students will feel compelled to use such tools.
Parents face a similar dilemma. Even if they worry about the long-term effects of AI on their children’s cognitive development, they may feel pressured to subscribe to expensive AI services to keep their children competitive. The structure of the education system, which emphasizes test performance and college entrance exams, leaves little room for individual choice. If AI is layered onto a system already criticized for rote learning, excessive memorization and an obsession with correct answers, the result could be a further narrowing of thinking and a decline in overall intellectual diversity.
Efforts to train AI talent, establish specialized schools and develop teacher capacity are all necessary. However, public resources are limited and priorities must be carefully set. It is worth asking whether adding AI-related initiatives on top of an already strained education system truly benefits future generations. When policies are driven by ambitious targets and timelines, there is a risk that schools focus on meeting measurable indicators rather than fostering meaningful learning.
For AI to function as a constructive “training wheel” in education, certain conditions must be met. Qualified teachers must guide its use, parents need education on its impact and evaluation systems must be reformed. The ability to quickly find correct answers should no longer be considered valuable in itself.
Instead, education should aim to preserve curiosity — the impulse to question what lies beyond the polished answers generated by AI.
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.
댓글
0