Human and non-human actors are increasingly interdependent in contemporary society. Generative artificial intelligence(AI), exemplified by ChatGPT, is now reshaping how people perceive and act through its language-based interactivity, and education is...
Human and non-human actors are increasingly interdependent in contemporary society. Generative artificial intelligence(AI), exemplified by ChatGPT, is now reshaping how people perceive and act through its language-based interactivity, and education is undergoing a similar AI-driven transformation. The benefits of such technologies, however, depend critically on the perceptions of their users; teachers, as the principal agents who apply AI in instructional design, assessment, feedback, and counseling, play a pivotal role. This study therefore explored elementary-school teachers’ perceptions of generative AI, traced the processes by which they designed instruction with it, and identified how those perceptions influenced design outcomes. Therefore, the research questions are as follows: First, What perceptions do elementary school teachers hold about generative AI? Second, Through what processes do elementary school teachers design instructions using generative AI? Third, How do elementary school teachers’ perceptions of generative AI influence their instructional design?
In this qualitative study, participating teachers engaged in classroom scenario-based activities, collaboratively designing lesson plans by interacting with generative AI, ChatGPT. Researcher collected data through direct observations and semi-structured interviews, applying coding and thematic analysis. Teachers’ perception toward generative AI were categorized into four distinct approaches: assistance-acceptive, tool-oriented, power-acceptive, peer-collaborative. Teachers identified as assistance-acceptive frequently rejected AI suggestions and evaluated them with skepticism. Those in the tool-oriented group routinely cross-referenced AI output with textbook resources, ultimately deferring to human judgment. Authority-acceptive teachers mostly accepted AI suggestions without substantial revisions. In contrast, the peer-collaborative group engaged in cooperative, reciprocal interactions with the generative AI, approaching it as a collaborative partner.
A qualitative analysis revealed that the instructional design produced by peer-collaborative teachers received the highest average scores, demonstrating a statistically significant difference compared to those created by assistance-acceptive teachers. These findings highlight that perceiving generative AI as an active collaborator and engaging in mutually complementary interactions can substantially enhance instructional design quality. This provides important insights into establishing sustainable integration of generative AI and fostering effective human-AI collaboration within various educational environments.