International Journal of
Information and Education Technology

Abstract

This study examines Preservice Teachers’ (PSTs) traditional lesson-planning practices, their proficiency levels, their interaction patterns with a specialized Artificial Intelligence (AI) powered lesson planning tool called Madlen, and their experiences and reflections throughout this integration process using a qualitative case study design. Data were collected through semi-structured interviews, an assessment rubric for AI supported lesson planning, and chat transcripts documenting PSTs’ interactions with the AI tool. Ten PSTs from primary education and Turkish language education programs participated in the study, each designing three lesson plans with Madlen following structured AI literacy training. Baseline analyses revealed that PSTs possessed structured, outcome-oriented planning competencies. However, interaction transcript analyses showed that most participants employed single turn, product-oriented prompts with minimal iterative refinement or critical evaluation of AI outputs. Although participants acknowledged the potential of AI tools to enhance time efficiency, content diversity, and creative lesson design, their actual engagement patterns indicated that these positive perceptions did not translate into deep, critically reflective AI use. This perception-action gap, in which declarative knowledge about effective AI use coexists with superficial interaction behaviors, carries significant implications for assessing AI literacy in teacher education contexts. These findings suggest that AI literacy training alone is insufficient; PSTs require scaffolded opportunities to develop skills in iterative prompting, critical evaluation of outputs, and pedagogical adaptation. This study contributes to the growing literature on AI in teacher education by providing direct evidence from human AI interaction data, moving beyond perception-based approaches that dominate current research.