Evaluation of the AI-Teacher Teaching Tasks Spectrum via Practitioner Review

This pilot study aims to evaluate the AI-Teacher Teaching Task Spectrum (AITTTS), a framework designed to categorise human-AI intervention levels based on the teaching tasks and their suitability for AI or human intervention. The primary objective was to provide preliminary validation of the framework’s practical utility by examining its alignment with current literature and gathering practitioner feedback. A systematic literature review was conducted, focusing on three key studies that offered insights into AI-teacher task delegation. Additionally, a structured survey was used to collect data from three expert practitioners in AI and education, with Fleiss’ Kappa applied to measure agreement. The findings indicated substantial agreement (Fleiss’ κ = 0.73) on the framework’s validity, particularly for identifying tasks suitable for AI, such as procedural and knowledge-based activities. However, the study’s small sample size and limited geographic diversity restrict the generalisability of the findings. Disagreements, especially regarding AI’s role in creative and relational tasks, highlight areas requiring further exploration. As the first step in an iterative research agenda, this pilot study provides foundational insights into the framework’s potential. Future research will involve expanded participant samples, broader educational contexts, and iterative refinements to enhance the framework’s applicability and generalisability across varied educational settings.