<![CDATA[Recent advances in generative artificial intelligence (AI) have introduced new opportunities for supporting teaching and assessment in science education. This study aims to examine teachers’ perceptions of the use of ChatGPT in supporting science epistemic assessment. A quantitative survey was conducted with 119 junior high school science teachers in Cimahi, Indonesia, using a structured questionnaire analyzed through descriptive statistics. The findings indicate that teachers generally perceive ChatGPT as a useful tool for instructional preparation, particularly in developing teaching materials and organizing lesson content. However, its integration into classroom interaction and epistemic assessment practices remains limited, especially for more complex and high-stakes tasks such as evaluating students’ reasoning and assisting with grading. While teachers demonstrate a strong understanding of the epistemic nature of science, this understanding is not yet fully reflected in the use of structured assessment practices supported by AI. These findings reveal a gap between the perceived potential of generative AI and its practical implementation in epistemic assessment, highlighting the need for pedagogical frameworks and validated AI-supported assessment tools to support students’ scientific reasoning and epistemic engagement]]>
Copyrights © 2026
