ChatGPT for Science Lesson Planning: An Exploratory Study Based on Pedagogical Content Knowledge

Abstract

Contemporary education is evolving in a landscape shaped by technological advancements, with generative artificial intelligence (AI) gaining significant attention from educators and researchers. ChatGPT, in particular, has been recognized for its potential to revolutionize teachers’ tasks, such as lesson planning. However, its effectiveness in designing science lesson plans aligned with the research-based recommendations of the Science Education literature remains in its infancy. This exploratory study seeks to address this gap by examining ChatGPT-assisted lesson planning for primary schools through the lens of a sound theoretical framework in Science Education: pedagogical content knowledge (PCK). Guided by the question, “What are the characteristics of lesson plans created by ChatGPT in terms of PCK?”, we designed four interactions with ChatGPT-4o using carefully constructed prompts informed by specific PCK aspects and prompt engineering strategies. Using qualitative content analysis, we analyzed data from these interactions. Findings indicate that incorporating PCK elements into prompts, using layer prompting strategies, and providing reference texts to ChatGPT might enhance the quality of AI-generated lesson plans. However, challenges were identified. This study concludes with guidelines for the teacher–ChatGPT co-design of lesson plans based on PCK.

Ähnliche Arbeiten