Designing Chatbots to Promote Systems Thinking in Environmental Science Education

• In Event: AERA Poster Session Friday 9:45am
  In Poster Session: STEM Education and Computational Learning Innovations

Abstract

Understanding the climate system requires systems thinking to consider intersecting components and relationships. Modeling can enhance systems thinking. We designed chatbots powered by large language models (LLMs) to provide modeling feedback and highlight multifaceted systems perspectives. We piloted the chatbots in a two-week environmental science curriculum. Sixty-one high school students created models in groups and interacted with the chatbots to revise models. We find that students incorporated more system components and complex causal ideas following chatbot interactions. However, students’ inquiries and resulting models remained focused on linear links, with limited attention to feedback structures. We discuss the potential of emerging technology to deepen systems thinking and the need for science instruction that scaffolds complex causal ideas and interactions with LLMs.

Authors

• Ha Nguyen, University of North Carolina - Chapel Hill

• Erick Valdez, Orange County Department of Education

• Jennifer Pei, University of California - Irvine

• Sara Ludovise, Orange County Department of Education

• Rossella Santagata, University of California - Irvine