Perceptions and Intentions to Use Generative Artificial Intelligence Among First-Year Medical Students in Japan: Cross-Sectional Survey Study (Preprint)

Abstract

<sec> <title>BACKGROUND</title> The rise of generative artificial intelligence (gAI) has created both opportunities and challenges in higher education. Although the potential benefits of learning support are widely recognized, little is known about how incoming medical students in Japan perceive and intend to use such technology. </sec> <sec> <title>OBJECTIVE</title> This study investigated the status of gAI usage, learning behaviors, and perceptions of first-year medical students in Japan. </sec> <sec> <title>METHODS</title> An anonymous online survey was conducted among 118 first-year medical students at Chiba University in April 2025. The questionnaire assessed prior gAI use, willingness to learn, perceptions of gAI, and the intention to use it academically. Likert scales, correlation analyses, and content analyses of free-text responses were used. </sec> <sec> <title>RESULTS</title> Of the respondents, 84.7% had prior experience with the gAI, primarily in language learning and information gathering. However, only 49.2% had learning experiences, mostly through informal sources, such as web browsing and peer interaction. Students showed a high willingness to learn about gAI (mean score: 4.3/5.0), which correlated with positive perceptions. Despite this interest, attitudes toward using gAI for academic assignments were neutral (mean 3.0/5.0). Content analysis of the open-ended responses revealed three types of attitudes: positive, cautious, and negative. </sec> <sec> <title>CONCLUSIONS</title> Although most students used the gAI, their limited exposure to formal learning suggests that self-directed experience alone may not foster confidence or informed use. Neutral attitudes and mixed qualitative responses highlighted the need for structured gAI literacy education that balances the benefits of ethical and critical considerations in medical education. </sec>

Ähnliche Arbeiten