This study aims to identify the types and characteristics of prompts generatedby learners, the characteristics of thinking and exploration activities appearing inthe prompt generation process, and the effects of generative AI usage experienceon learne...
This study aims to identify the types and characteristics of prompts generatedby learners, the characteristics of thinking and exploration activities appearing inthe prompt generation process, and the effects of generative AI usage experienceon learners' prompt generation ability and self-efficacy by analyzing promptsgenerated by elementary school students using the generative AI 'Wrtn'. Theparticipants were 26 fifth-grade elementary school students in Busan, whoperformed learning activities through a six-session program consisting of creation,information explanation, and problem-solving activities. A total of 819 collectedprompts were analyzed by type, and quantitative and qualitative results weresynthesized by conducting pre- and post-surveys and teacher focus groupinterviews (FGI).
The results of the study are as follows. First, the prompts generated by learnersappeared in the order of Information Request (33.34%), Creation (30.65%),Elaboration (14.78%), Format Request and Revision (14.66%), Emotional Expression(4.27%), Judgment and Evaluation (1.95%), Prediction and Application (0.37%), and Comparison and Contrast (0%). This indicates that students actively engaged inactivities to proactively create, modify, and refine outputs, going beyond simpleinformation seeking. Second, in the prompt generation process, students realizedconcrete self-regulated learning processes through 'self-reaction' and 'active learning' by critically evaluating and re-requesting AI responses instead of simplyaccepting them. Third, the survey analysis confirmed that repetitive masteryexperiences provided by AI significantly improved students' self-efficacy regardingquestion generation and problem-solving.