The Discerning Principal: What Prompting-Uptake Behaviors Characterize High Performers in Human-AI Collaboration?
Hui Liu, Y. Xu, Beichen Hu, and 1 more author
Education and Information Technologies, 2026
Under Review
As Generative AI (GAI) reshapes learning, exploring interaction behaviors is crucial for enhancing human-AI collaboration. This study uses Random Forest analysis to examine the behaviors of high vs. low performers (n=24) in an innovative instructional design task[cite: 10, 11, 56, 57]. Results characterize high performers as "discerning principals" with five core traits: intentionality, expertise-based judgment, strategic rejection, reflective adaptation, and systemic orchestration. These findings provide pedagogical insights for restructuring human-AI interaction frameworks[cite: 57].