Enhancing student–AI interactions: The 5W1H scaffold in a 5E model

Despite the growing interest in generative AI (GAI) in education, limited research has investigated how a 5W1H questioning scaffold supports effective student–GAI interactions within the 5E model and how such interactions shape learning outcomes. This study used design-based research involving 19 students across five iterative rounds. Across the rounds, the intervention was refined using evidence from academic performance, questionnaires, and interviews, with iterative adjustments to both the AI tools and the 5W1H scaffold. The findings indicated that the students became increasingly capable of using the 5W1H scaffold to interact productively with the GAI tool. Compared with their initial performance, they also showed improvements in learning interest, knowledge acquisition, and academic achievement. The interview data suggested generally positive perceptions of the interaction among AI, peers, and the instructor in the GAI-supported 5E environment, whereas the interview data suggested that the anthropomorphic features of the GAI tool were perceived positively and may have supported the students’ acceptance of AI. This study offers practical implications for designing GAI-supported scaffolds to foster interdisciplinary inquiry in 5E-based instruction.