Cha Sang -kyun, Stanford University Human -oriented AI Research Institute Scholar Fellow, Seoul National University Honorary Professor
The warning that artificial intelligence (AI) will replace human jobs has long been continued. In particular, Professor Jeffrey Hinton, a pioneer in Deep Learning, who won the Nobel Prize in Physics in 2024, predicted that in 2016, AI will replace a radiologist in five years. But the reality of 2025 is different. The world still needs a radiologist. However, the difference is that a doctor who uses AI replaces a doctor who does not use AI. AI is not a human job, but a tool used to create a new world by amplifying human abilities. Knowing the limits of AI and finding the value of humanity will be the starting point for discussions on education in the AI era.
Feeding from a uniform simple memorization
You have to develop the ability to solve the real problem
Humanity that AI cannot also need to culture
The state should provide a platform for experimentation
In the AI era, the world is filled with a network of people, AI agents and human noids. People are in charge of unmatched work and legal responsibility without a fixed manual. AI agents learn from accumulated data. Changes in the labor market are also inevitable. In response to this change, the nation that leads the educational innovation will lead the world. If the current SAT test and the university entrance examination system, which dominates Korea’s elementary, middle and high schools, will be maintained, it will be poisonous in the future AI era.
In the field of education, there are many voices that are afraid of early education. The representative logic is that the experience of ‘Halluscination’, which AI gives a fake answer plausible, confuses children. But you should think backwards. The Harlucity experience itself can be educated. The ability to filter out wrong information through critical thinking, ethical thinking, source verification, and collaboration is essential for future society.
In linguistics, human intelligence is viewed as the sum of ‘compete’ and ‘performance’. Competances are inherent abilities, such as knowledge or algorithms of the world. In this area, AI already surpasses humans. But how creatively expressing its ability in the performance, that is, in the real context, is still up to humans. Education should be redesigned from this view. For example, instead of interpretation of the SAT test, it is necessary to turn into language education where students are directly on stage and become players to present their opinions and write writings in the form of theater. In the age of AI’s fluent in English and Korean, we must consider what the difference between English and Korean education is.
The same is true of history, society, mathematics, and science education. It should be taken away from the way you memorize the description of the textbook, and to use the actual data to an education that attempts to interpret various interpretations. For example, if you analyze economic indicators, demographic statistics, elections, and environmental changes and conduct discussions and projects based on it, students will experience the subject as a living reality.
One of the things that should be emphasized in the AI era is a collective experience. Rather than staying in a virtual space alone, learning on the premise of cooperation, such as sports -like team play, ensemble and theater, is a soil that develops humanity. At the same time, students need to use AI as leverage to expand their thoughts and create creative outputs. In an evolving environment as humans and AI interact, the essence of future education is to consider and express ‘human performance that AI cannot’.
The most important thing is the philosophy of education policy. So far, Korean education has been a way for the government to design and descend everything. The ambitious digital AI textbook policy, which began, also increased the fatigue of the field due to lack of autonomy and social consensus between teachers and students. In the AI era, this uniformity is rather poisonous. Now, we should recommend various educational experiments centered on teachers and students, and establish an ‘open source education ecosystem’ that shares its achievements. Educational content is made in the classroom, and it is desirable to develop as it refers to the community again.
The world’s major AI companies are also struggling with their own strategies in the field of education. Open AI provides tools that can be used by teachers and students through the chat GPT and Copilot service, and is laying the foundation for ‘AI textbooks’ in cooperation with Khan Academy. However, the closure and cost structure of the model is a constraint of the education site. Google provides an environment where teachers and students can directly test, modify and utilize AI through platforms such as open source models and Google Class Room. This helps learners grow into a subjective learner with AI production and utilization, not just consumers. Open AI’s user -friendliness, Google’s open source scalable philosophy is a different way but complementary. Rather than relying on a model of a particular company, Korean education should move to the direction of strengthening AI’s ability, creativity, and ethics by balancing these strategies.
What alternatives will Korean Sovereign AI offer? The role of the state is not to give the right answer, but to recommend various experiments and establish an open platform for experiments. The AI learning module and class cases voluntarily created by schools, teachers, and students are needed nationwide and further shared and developed worldwide. This will be as if open source software has become an engine of global innovation, the open source education ecosystem will make Korean education a global leading model.
Cha Sang -kyun, Stanford University Human -oriented AI Research Institute Scholar Fellow, Seoul National University Honorary Professor
