"When I called his name / he came to me / and became a flower."
Every March, Kim Chun-soo's poem "Flower" becomes relevant again. Not simply because it is the season when flowers bloom. It resonates because March aligns with the human process of assigning meaning to new objects or situations, questioning them, and ultimately understanding them.
Unfamiliar worlds do not easily acquire meaning. Among all that passes by, meaning emerges only after the laborious process of deciding what to pause and observe, what to name, and what to accept into oneself. The poem "Flower" represents the sophisticated form of thinking that humans have long maintained—the act of grasping the unfamiliar until it finally becomes part of one's own understanding, rather than simply accepting or consuming the world as given.
But we now live in an age of impoverished thinking. With the spread of generative artificial intelligence, accepting AI's instant outputs is becoming the new standard for acquiring knowledge, replacing the time-consuming work of inquiry, logic-building, and judgment. Before I can call something by name myself, AI produces tens of thousands of meaningless flowers in mere seconds.
While it is problematic that adults avoid the effort of intellectual labor, the crisis in education—where the next generation is being raised—is more alarming. The process of thinking is visibly weakening in classrooms today.
According to a research report titled "Survey on the Educational Use of Generative AI and Related Needs" published in January by the Seoul Metropolitan Office of Education's Education Research and Information Institute, 94.7% (25,104) of 26,531 students surveyed in Seoul had used generative AI. However, 93.4% (3,121) of 3,344 teachers surveyed expressed concern about students' excessive dependence on AI, while 92.4% (3,089) worried about acceptance of biased information. Teachers concerned about uncritical acceptance of information reached 92.5% (3,096).
The situation is similar in other countries. According to joint research released early this year by the American Association of Colleges and Universities (AAC&U) and Elon University, 90% of 1,057 U.S. college faculty members surveyed expressed concern about weakening critical thinking skills, while 83% worried about declining attention spans.
This is by no means an argument to vilify AI as bad technology. AI has already permeated education, industry, research, and administration. The issue lies in usage standards and methods—more precisely, the question of what must be cultivated first in order to use AI properly.
To utilize AI effectively, one must first possess the ability to think deeply, ask precisely, withhold easy belief, and question and discern to the end. Deciding what questions to ask and identifying errors and biases in the answers returned are ultimately human responsibilities.
More important than the skill of crafting prompts proficiently is the literacy, background knowledge, and logical discernment needed to evaluate AI-generated explanations. When people with weak foundational thinking, literacy, and discernment rely on AI, they accept information uncritically rather than reading selectively.
The current crisis cannot be viewed merely as a matter of individual diligence. Responsibility lies in broader quarters. At the level of national destiny, we cannot simply proclaim "AI First." Clearer standards for AI use and evaluation principles must be established in curricula. Methods that allow verification of the thinking process must be introduced.
The Ministry of Education recently prepared AI ethics guidelines for universities, emphasizing principles of academic integrity, human-centeredness, transparency, and reliability—a move not unrelated to these concerns. Education must center on "how properly one thought and judged" rather than "how quickly one found the answer."
I recite "Flower" once more. To call something by name means to have understood it in one's own way through deep thinking. Time spent observing, questioning, doubting, and judging is necessary before meaning can emerge. AI may shorten that time, but it cannot do that work in our place.
Equipping the next generation with the ability to think, question, and discern for themselves—even in the face of tremendous technology—is what education must do today.
