AI-Supported Pedagogical Supervision: A Theory-Building Framework for Understanding Feedback, Cognitive Processing, Reflective Practice and Pedagogical Decision-Making

The increasing integration of generative artificial intelligence (AI) into teacher education and pedagogical supervision requires explanatory frameworks capable of clarifying how AI-generated feedback may support professional learning processes. Existing research has predominantly focused on technological adoption, implementation challenges, and user perceptions, while comparatively limited attention has been devoted to the cognitive and reflective mechanisms involved in AI-supported pedagogical supervision. In response to this gap, this article proposes a theory-building conceptual framework explaining how AI-supported pedagogical supervision may influence pedagogical decision-making through sequential mechanisms involving feedback quality, cognitive processing, and reflective practice. Drawing on feedback theory, Cognitive Load Theory, reflective practice literature, and distributed cognition perspectives, the proposed framework conceptualises AI not as a direct instructional agent, but as a support system embedded within professional pedagogical reasoning processes. To facilitate future empirical investigation, the article proposes a validation framework based on covariance-based Structural Equation Modelling (CB-SEM). This methodological specification is intended solely as a research agenda for subsequent studies and does not constitute empirical testing of the model. As a conceptual contribution, the article advances a theoretically integrated explanation of how AI-generated feedback may influence professional learning processes. By articulating feedback quality, cognitive processing, reflective practice, and pedagogical decision-making within a coherent framework, it offers a foundation for future empirical research and theory development in AI-supported pedagogical supervision.