Kason Ka Ching Cheung, Department of Curriculum and Instruction, The Education University of Hong Kong, China
When ChatGPT is first introduced into the field of science education, educators asked questions like “Does generative artificial intelligence (GenAI) hamper students’ learning of nature of science?” and “Does GenAI simplify the works of scientists?”. To begin with, I will share some voices from frontline science teachers who are concerned about the incorporation of GenAI technologies in science classrooms according to my teacher training experiences. I will present findings from two systematic reviews related to the use of AI and GenAI in science education respectively. From these reviews, I will show that current instruments and teaching interventions in studies that use GenAI to facilitate science learning do not consider disciplinary learning in science and nature of science. I argue for the need for reconceptualising nature of science under the era of GenAI. More importantly, I will also discuss some research findings across K-12 and higher education on students’ understanding of how GenAI can be applied to scientific activities. There are two salient findings: students do not develop a sufficient understanding of how GenAI creates scientific information; students recognise the aims and values and source of knowledge of GenAI in communicating scientific information, yet recognition of these aspects is disconnected from social-institutional aspects. Therefore, I call for the need for revising the pedagogical framework of GenAI that considers nature of science education as well as facilitating scientific activities. I also highlight some suggestions that guide future research directions of GenAI in science education, and aligning with the views of in-service science educators.