Deliberative Instructional Agents: transforming professional learning with the responsible integration of Artificial Intelligence

Inquiry is increasingly recognised as a transformative approach to professional learning. It encourages educators to critically reflect on classroom practice by engaging with evidence of practice, educational research, and their own experiential knowledge. Despite its promise, the workload demands of teaching make sustained engagement in Professional Inquiry difficult to maintain. This paper introduces Deliberative Instructional Agents (DIAs), a new class of educational technology that leverages advances in Artificial Intelligence (AI) to make Professional Inquiry feasible within the realities of everyday teaching. A DIA can efficiently capture evidence of classroom practice and automate the use of relevant research to quickly render learning analytics. DIAs then encourage focused and timely reflection on critical instructional moments by offering teachers the opportunity for text-based interaction with a specialised language model, engineered to perform like an instructional coach. This paper presents a proof-of-concept DIA designed to support professional learning in dialogic pedagogy – a demanding instructional approach centred on purposeful discussion, reasoning, and debate. Five design principles are proposed for the responsible development of AI-powered technology to support professional learning. These principles are intended to guide developers and help educators navigate what is becoming a crowded marketplace of new technology.