Exploring teacher adoption of AI: A structural analysis of Microsoft Copilot in education

Abstract

This study aims to investigate the driving factors influencing teachers’ intention to adopt an AI-powered assistant (Microsoft Copilot) in their professional development.  This study attempted to validate ten hypothetical assumptions derived from notable theoretical models (UTAUT and TPACK). A survey was conducted with 280 teachers, who responded through Google Forms. The data were analyzed using Generalized Structured Component Analysis (GSCA) to test the proposed research model. The findings showed that both Technological Pedagogical Knowledge (TPK) and Perceived Ease of Use (PEOU) had a significant impact on teachers’ self-efficacy and their intention to use Copilot. PEOU also played a key role in influencing Perceived Usefulness (PU) and Teaching Self-Efficacy (SE) while TPK directly affected PU and Behavioral Intention (BI). Interestingly, and somewhat unexpectedly, PU did not show a meaningful influence on either SE or BI. These results suggest that how easy a tool is to use and how well it fits into teachers’ existing pedagogical knowledge may matter more than how useful it appears on the surface. Our proposed model explains 55.9% of the variation in the data. The study’s findings are expected to make important contributions to the academic and practical aspects of applying AI to enhance teachers’ digital competence.