Abstract
<p>The proliferation of generative AI in education has fundamentally unsettled teachers' professional identities, yet little is known about how this identity transformation unfolds or how it differs between novice and experienced teachers. Drawing on a qualitative phenomenological design, this study investigated 138 teachers' (64 novice, 74 experienced) identity transformation through four distinct cohorts (two novice cohorts, n=64; two experienced cohorts, n=74), each comprising four weekly workshops (totalling 16 workshop sessions). Observational data were collected by two observers per session, yielding 32 observation records grounded in the Integrated AI Triad (IAT) model. Analysis of reflective journals, semi structured interviews, observation notes, teacher designed tasks, and retrospective self assessment surveys revealed four stages of identity transformation—Awareness, Experimentation, Integration, and Transformation—with differential trajectories by career stage. Novice teachers progressed faster through early stages but achieved shallower transformation, with technical fluency outpacing pedagogical depth. Experienced teachers exhibited slower initial progress but deeper, more sustained transformation, culminating in a coherent Pedagogical Prompt Engineer identity. The emergence of Reverse PCK—a recursive process wherein teachers identify AI's limitations and strategically redesign pedagogical content knowledge—was a key mechanism across both groups. Findings extend TPACK and teacher identity theory by introducing two novel constructs: Reverse PCK as a pedagogical mechanism and the Pedagogical Prompt Engineer as an emergent professional identity. Implications suggest differentiated professional development that respects novice and experienced teachers' distinct starting points, cultivates critical AI evaluation, and positions teachers as strategic designers who preserve human agency in AI mediated learning environments.</p>