Title: “Artificial Intelligence (AI) and Science Education Pedagogy”

Abstract: Artificial Intelligence (AI) offers promise for strengthening science education by extending the practices of experienced teachers and exploring novel opportunities. Our partnership including teachers, computer scientists, and learning scientists has developed and refined AI solutions for science education guided by Knowledge Integration (KI) pedagogy. Initially, the partners designed AI tools to guide students’ written explanations. Students using the tools were more likely to revise their explanations and improve their reasoning than students receiving typical guidance, especially if they followed suggestions to reconsider the implications of the interactive models in the curriculum. Subsequently, we designed AI tools to identify the ideas in students’ explanations and collaborated with teachers to figure out which student ideas to select. Guiding on ideas resulted in further increases in the proportion of students who revised. This talk illustrates ways that AI can transform science education and how exploring uses of AI spurs refinement of KI pedagogy. The talk concludes with a vision for future roles for AI in education.

講題：“The Impact of AI on Science Education Practices: Challenges and Possibilities”

摘要：This talk will begin with an overview of key issues regarding the potential impact of Artificial Intelligence (AI) on science education practices, based on current research and perspectives on how AI may transform science and school practices. Specifically, research highlights the importance of developing students' critical thinking and epistemic justification to evaluate AI-generated text. Our preliminary results on engaging students in using a generative AI tool as their collaborative AI peer to build scientific models will be reported and discussed. Additionally, we have examined how students interacted with the AI tool and whether they employed epistemic justification to evaluate AI-generated text, as well as the relationship between these interactions and students' science learning and achievements, which will also be presented in this talk. The goal of this talk is to encourage discussion and collaboration on future research that advances our understanding of how to design AI-infused science learning environments to best benefit students and address important goals in science education.