Shifting Ethical Concerns about Generative AI among First-Year University Students: A Comparative Text Analysis from 2024 and 2025

Abstract

As generative AI becomes embedded in education, understanding students’ ethical concerns is essential for developing responsive AI literacy programs. This study analyzes free-text responses from first-year university students in Japan in 2024 and 2025, focusing on negative impressions of generative AI. A coding framework of nine ethical categories was developed based on international and national guidelines, and responses were analyzed using KH Coder. Notably, concerns about employment decreased significantly in 2025, while worries about privacy, misuse, and copyright persisted. Abstract principles such as fairness or accountability were rarely mentioned. These results underscore the importance of supporting students in articulating foundational ethical concepts. Future work will investigate the relationship between students' ethical perceptions and their actual engagement with AI tools in learning environments.