Toolprint and Cognitive Depth: A Comparative Analysis of Novice and Experienced Teachers' AI-Generated Challenging Questions within the IAT Framework

<p>The integration of generative AI into education demands systematic investigation of how teachers interact with AI tools. This exploratory study introduces Toolprint—the linguistic, structural, and cognitive signatures that AI tools imprint on generated content—operationalized through Interrogative Density (questions per 100 words) and Interrogative Range (ratio of higher-order to lower-order questions based on Bloom's Revised Taxonomy). Analyzing 247 challenging questions from 89 Enriched Content samples generated by 38 novice and 51 experienced Iranian teachers within the IAT framework, we found that experienced teachers produced significantly higher proportions of higher-order questions (66.9% vs. 41.7%), greater Interrogative Density (1.49 vs. 1.29), and a threefold higher Interrogative Range (2.03 vs. 0.71). Qualitative linguistic analysis revealed that experienced teachers' questions were more open-ended, employed higher-order verbs more frequently, and embedded more real-world connections. While the exploratory design limits causal attribution of stylistic features to specific AI tools, the findings suggest that teacher professional experience—particularly pedagogical content knowledge—substantially influences the cognitive quality of AI-generated content. The study proposes Toolprint Recognition as a dimension of teacher AI competence within the IAT framework, offering both conceptual vocabulary and empirical evidence for designing professional development. However, controlled studies with systematic tool assignment are needed to establish stable Toolprint signatures and causal relationships between tool characteristics and output quality.</p>