The generative AI ethics landscape as seen by Chinese middle school students

With the rapid integration of Generative AI in education, understanding students' ethical perspectives is crucial for effective AI ethics education. Five hundred and ninety four middle school students' agreement levels on five AI ethical principles (beneficence, non-maleficence, justice, autonomy, explicability) adapted from previous research, and the rationales underlying their choices were investigated using a questionnaire. Results showed that students expressed the highest agreement with “beneficence” and “autonomy,” though overall responses leaned toward neutrality. Independent AI use and family discussions predicted higher agreement; urban-rural differences were non-significant. Qualitative analysis identified themes in students' ethical reasoning. These findings offer evidence-based guidance for adolescent AI ethics education.