Teachers’ Anxiety and Students’ Expectations Toward Generative AI in Education: Evidence From a Two-Wave National Survey in Japan (2024–2025)

Author Information

Nagayuki Saito, Sendai University, Japan
Chiaki Hashimoto, Sendai University, Japan
Yasumasa Yamaguchi, Sendai University, Japan

Abstract

Generative AI is increasingly being adopted in education for lesson preparation, instructional material development, and feedback support, while also raising concerns about academic integrity, fairness in assessment, and data protection. This study examines the structure of teachers’ and students’ expectations (perceived benefits) and anxieties (perceived risks) regarding pedagogical uses of generative AI, and how these perceptions changed over time. We conducted two nationwide repeated cross-sectional online surveys in Japan in March 2024 and March 2025. In both waves, ten benefit items and ten risk items were measured on five-point Likert scales, and differences between teachers and students as well as year-to-year changes were compared. Perceived benefits were generally high in both groups, but students consistently reported higher levels than teachers, with a larger increase from 2024 to 2025. In contrast, teachers reported relatively higher concerns in governance-related domains, including output accuracy and reliability, cheating and plagiarism, privacy, and the risk of being wrongly suspected of misconduct, whereas students expressed comparatively stronger concerns about face-to-face learning and the changing role of teachers. Overall, the findings suggest that the diffusion of generative AI in education is characterized by the parallel salience of perceived benefits and anxieties rather than a simple trade-off.

Save as a Favourite