ABSTRACT Background Generative artificial intelligence (GenAI) tools, such as ChatGPT, present both opportunities and concerns in education. They can support collaborative knowledge construction by providing systematic problem‐solving support. However, a strong focus on critical thinking and validity judgement is needed. GenAI's role in collaborative learning remains underexplored, especially in fostering critical thinking. Objectives This study examines GenAI's role in a collaborative task where 75 pre‐service teachers co‐create lesson plans on sustainable development. Methods Process mining is used to identify sequential patterns that indicate critical thinking. These patterns draw on video analysis and on categorising collaborative groups into high and low performers based on lesson‐plan quality. We analyse different prompt types (conceptual, pedagogical, and refining prompts). We also analyse activities, such as writing, editing, searching for other online sources, and copy‐pasting information. The sequential interplay of these activities reveals important aspects of critical engagement with GenAI. Results and Conclusions The results showed that participants frequently engaged in an iterative process of prompt refinement, indicating reflective use of AI. They also integrated conceptual and pedagogical aspects in their prompting. High‐performing groups demonstrated deeper engagement in activities such as iterative writing and corroborating GenAI content with other online resources. In contrast, low‐performing groups showed repetitive prompting and reliance on AI‐generated content. The results illustrate GenAI's role in collaborative and reflective learning and call for targeted interventions to support critical thinking. Educational interventions should focus, for example, on prompt refinement, guided reflection, and verification of GenAI information against other sources. Theoretically and methodologically, our analyses provide a foundation for further research.
Building similarity graph...
Analyzing shared references across papers
Loading...
Auli Lehtinen
Faisal Channa
Piia Näykki
Journal of Computer Assisted Learning
University of Jyväskylä
Building similarity graph...
Analyzing shared references across papers
Loading...
Lehtinen et al. (Tue,) studied this question.
www.synapsesocial.com/papers/69f6e60f8071d4f1bdfc6a40 — DOI: https://doi.org/10.1002/jcal.70256
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: