Exploring Elementary Pre-Service Teachers’ AI Literacy in ChatGPT-Assisted Science Lesson Design
Yoon-Sung Choi, Min-Cheol KimThis study examined elementary pre-service teachers’ artificial intelligence (AI) literacy in ChatGPT-4o-assisted science lesson design. Eight third-year pre-service teachers at a national university of education in Korea participated in a voluntary, non-credit, ten-week program focused on elementary Earth science lessons and used GPT-4o through the web-based ChatGPT interface. Data included pre-task written reflections, ChatGPT interaction logs, lesson plans and instructional artifacts, reflective journals, observation notes, and final reflective reports. A framework-guided thematic analysis examined AI knowledge, AI skills, AI ethics, and AI attitudes, while participant-level trajectories and lesson-design artifacts were analyzed across data sources. Initial AI literacy was heterogeneous rather than uniformly limited. Five of the eight participants articulated some pattern-based, probabilistic, or context-sensitive understanding of ChatGPT, seven of the eight anticipated using prompts that specified instructional conditions, all eight recognized a need to verify AI-generated content, and initial attitudes ranged from affirmative willingness to predominant skepticism. During the program, six of the eight participants showed increasingly contextualized or iterative prompting, and seven of the eight demonstrated artifact-linked verification, revision, supplementation, or alternative tool selection. Attitudinal trajectories were characterized by conditional acceptance rather than a uniform increase in trust. Across the eight participant-level trajectories, the analysis identified refinement, continuity, limited change, and one episode-specific awareness–practice inconsistency. These findings indicate that ChatGPT-assisted science lesson design can provide a practice-based context for examining AI literacy as situated professional judgment involving scientific, curricular, pedagogical, and ethical decision-making, without implying a common developmental progression.