Comparing GenAI Summaries Versus Chatbots for Supporting Reflection in Healthcare Simulation
Abstract
Reflection is critical for learning. In healthcare, simulations help learners practice clinical and teamwork skills, while reflection supports improving practice. Learning analytics (LA) dashboards can support reflection by visualizing activity data, but their effectiveness is limited by learners’ ability to interpret data visualizations and domain knowledge. Generative AI (GenAI) may bridge this gap through narrative explanations of data. Yet little is known about how forms of GenAI support—such as chatbots or summaries—affect reflection. We conducted a comparative study with 41 nursing students who wrote reflections after a simulation, supported by an LA dashboard augmented with a GenAI summary or chatbot. Results suggest that written reflections differed across conditions: GenAI summaries help align reflections with learning goals, while chatbots encourage consideration of actions and plans. However, nursing teachers identified no difference in the quality of plans. These findings suggest GenAI configuration may shape how technologies support reflection.
// Source
Authors: Mikaela Elizabeth Milesi, Viktoria Pammer‐Schindler, Linxuan Zhao, Yueqiao Jin, Vanessa Echeverría, Abra Osborne, Riordan Alfredo, Jie Xiang Fan, Lixiang Yan, Dragan Gašević, Yi‐Shan Tsai, Roberto Martínez‐Maldonado
Institutions: University of Hong Kong, Monash University, Tsinghua University, Education University of Hong Kong, RMIT University, Graz University of Technology, Escuela Superior Politecnica del Litoral