PRE-SERVICE TEACHERS' PERCEPTIONS OF AI ETHICS AND ETHICAL DILEMMAS IN SCIENCE CLASSROOM CONTEXTS

Abstract

Pre-service teachers are increasingly responsible for guiding students through AI-related ethical challenges in science classrooms, yet little is known about how they perceive these challenges. In this study, 165 pre-service teachers completed a 19-item perception scale spanning five AI ethics domains and an engagement measure, and responded to eight open-ended dilemma scenarios set in science classrooms. Participants showed strong agreement on non-maleficence (M = 4.57) and privacy protection (M = 4.51) but lower and more varied responses on accountability (M = 3.73) and fairness (M = 3.94). Female participants scored significantly higher than male participants in overall perception, while no differences were found by major field or AI coursework experience. Scenario responses revealed that ethical concerns shifted depending on teaching context: learning integrity dominated in student assignment scenarios, verification responsibility in teacher use scenarios, and privacy-public interest tension in social application scenarios. Across scenarios, proposed solutions remained largely at the level of general prescriptions, such as ethics education and prior consent procedures, with few implementation-ready alternatives. These results point to the need for scenario-embedded AI ethics education within science teacher preparation programs. Keywords: AI ethics, ethical dilemma, pre-service teachers, science education, teacher preparation