Abstract
<jats:p>This study addresses a critical challenge in teacher education: the tendency of pre-service teachers’ reflective writing to remain largely descriptive, lacking the necessary intellectual depth, conceptual framing, and critical analysis. To investigate potential technological scaffolds, the purpose of this research was to examine the efficacy of ChatGPT as a conversational writing assistant designed to support and enhance the development of structured reflective thinking during initial professional preparation. Conducted among primary pre-service teachers at Iakob Gogebashvili Telavi State University, Georgia, in December 2024, the study employed a mixed-methods research approach. Data collection involved the systematic gathering and cross-validation of independent reflective essays, ChatGPT-assisted reflective essays, and comprehensive conversation logs captured directly from the AI interface. The empirical findings demonstrate that ChatGPT can significantly and positively impact the structural organization of reflective writing, expanding content parameters, and enhancing programmatic vocabulary use. However, the data concurrently reveals that AI integration can influence student writing in unintended ways, notably by diminishing the writer’s authentic voice through standardized phrasing. Consequently, the research underscores the urgent need to explore specific genres of prompt engineering and structured AI scaffolding that can effectively foster deeper metacognitive reflection without compromising original authorship. These insights offer valuable guidance for educational leaders aiming to integrate artificial intelligence responsibly into contemporary teacher training curricula.</jats:p>