The purpose of this study is to develop and verify the effectiveness of a prompt engineering education program incorporating question generation strategies, designed to enable elementary school students to utilize generative artificial intelligence (A...
The purpose of this study is to develop and verify the effectiveness of a prompt engineering education program incorporating question generation strategies, designed to enable elementary school students to utilize generative artificial intelligence (AI) correctly and effectively. With the rapid introduction of generative AI in educational settings, the need for ‘Generative AI Literacy’ education to ensure its correct and effective use is emerging. In particular, within the context of literacy education, ‘prompt design skills’—the ability to elicit desired answers from AI—and ‘metacognition’—the ability to regulate one’s own learning process—are being increasingly emphasized as core competencies. However, there is a lack of systematic education programs tailored to the level of elementary school students. In particular, as elementary school students face difficulties in structuring their thoughts into specific language for questioning, a systematic question generation support strategy is essential for effective prompting. Accordingly, this study designed an education program that utilizes question generation instructional elements as cognitive scaffolding to help students refine their thinking skills during the interaction process with AI.
To achieve these objectives, four core domains—‘Prompt Composition and Input’, ‘Analysis and Reflection on Generative AI Responses,’ ‘Prompt Refinement and Iteration,’ and ‘Prompting Inquiry and Utilization Attitude’—along with 12 constituent elements, were derived by analyzing the curriculum and prior research. Based on these findings, an educational content system was developed and subsequently validated for its effectiveness and appropriateness. Using these as design principles, the question generation instructional elements of Rosenshine et al.(1996)—such as procedural facilitators, modeling, guiding potential difficulties, and providing cue cards—were strategically placed at each instructional stage to alleviate students' difficulties in prompt writing. To verify the program's effectiveness, the program was implemented with upper elementary school students, and changes in generative AI literacy and metacognition were analyzed through a single-group pre- and post-test design.
The analysis results indicated that the education program had a significant impact on the students' competence development. First, students' generative AI literacy improved significantly after the education, with a very high effect size. Second, metacognitive abilities also significantly increased overall; in particular, ‘conditional knowledge,’ ‘planning,’ and ‘monitoring’ abilities among the sub-elements of metacognition showed significant improvement. However, no significant change was found in the ‘evaluation’ domain, which involves self-assessment after learning. Third, a positive correlation was confirmed between the improvement in generative AI literacy and metacognition. These results imply that the question generation instructional elements successfully functioned as scaffolding to support students' thinking. In other words, experiencing education in which prompts are written and refined based on question generation strategies shows that generative AI literacy and metacognitive abilities positively influence each other and improve together. Fourth, positive changes were also captured in the results of the prompts actually written by the students. Questions became more specific and clearer than before the education, and a process of improving output through a ‘plan-check-revise’ cycle was confirmed.
In conclusion, this study is significant in that it proved that prompt engineering education based on question generation strategies for elementary school students is an effective educational method for simultaneously fostering generative AI literacy and metacognition.