ChatGPT-generated socioscientific issues in the context of systems thinking: affordances and limitations

Abstract This study explores the potential of generative artificial intelligence (GenAI), specifically ChatGPT, in supporting the design of socioscientific issues (SSI) unit plans that foster systems thinking in high school science education. Employing document analysis, we systematically examined a series of outputs generated by ChatGPT 3.5 to produce an instructional unit focused on cross-state air pollution. Activities were coded using Stave and Hopper’s systems thinking taxonomy to assess the depth of student engagement with systems thinking. ChatGPT generated a structured SSI unit plan that provided opportunities for students to practice basic levels of systems thinking. However, GenAI outputs typically lacked specificity and contained inaccuracies. Iterative prompting led to incremental improvements but did not substantively elevate the complexity of systems thinking tasks. Teacher expertise and critical engagement remain essential for ensuring accuracy, depth, and pedagogical value.