Abstract:

This paper presents an ongoing design-based research project that investigates a novel, human-centred strategy for incorporating artificial intelligence (AI) into teacher professional development. Within this project, AI is conceptualised as a Pedagogical Mirror, and as a reflective artefact that challenges educators to critically analyse their own pedagogical reasoning and instructional decisions rather than as a tutor, assessor, or decision-maker.

The study addresses the growing need for ethically sound, thoughtful models of AI use in education, and it is set within the framework of in-service Cycle 2 and Cycle 3 public school teachers in Abu Dhabi. AI-generated analyses of teachers' own instructional artefacts are used as provocations for professional reflection, discourse, and sense-making rather than for assessing instruction quality.

The project employs a design-based research methodology that combines qualitative and mixed-method data collection with iterative protocol development. Teachers take part in a semi-structured reflective dialogue, submit teaching artefacts, evaluate an AI-generated "decision trace", and complete a post-reflection survey as part of an organised Pedagogical Mirror process. AI-generated decision trace reports, audio-recorded reflective interviews or focus groups, teacher-produced artefacts (lesson plans, brief classroom video segments, and written reflections), and a post-protocol survey that combines Likert-scale and open-ended items are the primary methods used to collect data. The AI output is explicitly descriptive and non-evaluative, and it is produced using a fixed, transparent prompting structure.

The project's three primary objectives are to:
(1) jointly develop and improve a Pedagogical Mirror protocol that incorporates procedural, ethical, and pedagogical safeguards;
(2) investigate how educators interact with, challenge, accept, or reject AI-generated accounts of their own practice; and
(3) investigate educators' opinions regarding the value, constraints, and professional ramifications of AI-mediated reflective inquiry.

This paper presents a novel methodological model for investigating human-AI interaction in professional learning contexts by outlining the conceptual framework, research design, and data collection strategy. Teachers' opinions, ethical concerns, and professional insights are captured through descriptive and content analysis of survey data, which is supplemented by qualitative thematic analysis of reflective dialogue transcripts using a hybrid inductive–deductive approach. By providing a scalable, context-sensitive approach to reflective AI integration that prioritises teacher agency, professional judgement, and ethical responsibility, the project is anticipated to influence future teacher education practices.

Keywords:

Artificial intelligence in education, teacher professional development, reflective practice, design-based research, human-centred AI.