Answers are grounded in this paper and cite their sources.
Creative problem-solving (CPS) is essential for addressing complex and uncertain problems, yet it is challenging for learners to develop high-quality solutions. Generative artificial intelligence (GenAI) offers new possibilities for supporting CPS by facilitating divergent and convergent thinking. However, how learners with different performance levels engage in human-GenAI collaborative CPS remains underexplored. This study investigated the CPS processes of 25 students in Grades 4-6. Specifically, descriptive statistics and lag sequential analysis were used to examine students' behavioral characteristics and sequences, respectively, while epistemic network analysis was employed to investigate interaction qualities. The findings indicated that primary school students exhibited more behaviors in the idea generation and solution validation stages, reflecting a task-oriented behavioral pattern supported by GenAI. Significant differences in behavioral sequences across performance levels were observed: high-performing groups (HPGs) demonstrated clearer CPS goals and more continuous, progressive behavioral sequences, whereas low-performing groups (LPGs) showed less defined goals and more fragmented behavioral sequences. In terms of interaction qualities, HPGs were more effective at facilitating CPS through strategic questioning and the productive use of feedback, while LPGs were characterized by more superficial interactions. These findings provided theoretical insights and empirical evidence to inform human-GenAI collaborative CPS in primary education.
No takes yet. Share an insight, caveat, or question.
Xu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: