[Paper Review] Generative AI Education — The Gap Between Design and Learner Response

Curricula utilizing generative AI are rapidly spreading from elementary through university levels. Yet empirical discussion of how best to use generative AI in education, and how it is actually experienced by learners, remains in an accumulation stage. Two studies address respectively: the gap between generative AI education in elementary/middle/high schools and students'' actual learning needs; and how generative AI use in university education operates as a matter of choice versus compulsion. Through these, the gap between the design intent of generative AI education and learner experience becomes more concretely apparent.

This paper explores the effect of generative AI on university students'' computational thinking and problem-solving ability. Comparing three groups — AI non-use, selective use, and mandatory use — through the required liberal arts course "Computer-Based Problem Solving" at J University, it empirically analyzes how AI can contribute to education in the digital age.

Educational Gaps and Demands at Elementary and Middle School Levels
At the elementary level, the gap between "protection" and "utilization" is the problem. School settings recognize students only as objects of protection, concentrating on personal information protection (69.2%) education. However students perceive AI as a friend and want concrete interaction methods like "how to speak to it (communication etiquette)," confirming the absence of practical education premised on safety. At middle school level, while critical thinking education for judging information reliability is the main focus, students had strong desire to utilize AI as more creative tools for music composition or image generation based on ethical judgment capabilities. This means education that opens creative opportunities alongside ethical regulation must run in parallel for middle school students.

The "Return to Basics" Phenomenon at High School Level
At high school, AI utilization is highest for exam preparation and career exploration, and school education has been conducted centered on creative activities. However, the more advanced utilization students experience, the more they paradoxically seem to harbor fundamental doubts: "Is my data safe?", "Is this information real?" The fact that high school students strongly demand personal information protection education (44.4%) suggests that as technical utilization capabilities advance, the thirst for accompanying risk management capabilities and ethical standards also grows.

Why We Need to Rethink Generative AI Education
Conclusively this research proposes redesigning the current AI curriculum through answers to "what do students want?" Resolving the mismatch between educational conditions and learning needs across all school levels is the most urgent task. Elementary students need friendly interaction training and basic etiquette; middle school students need a balance of verification capabilities and creative experience; high school students need advanced utilization skills alongside legal and ethical safety education enabling self-protection.