Data supporting the PhD dissertation: The Material Gaze - Reflective AI as a Design Practice

This repository contains research data accompanying the doctoral dissertation "The Material Gaze: Reflective AI as a Design Practice" (TU Delft, 2026). The research develops Reflective AI as a design practice that treats AI training processes as sites for designer self-reflection.<br>The repository includes four practice-based investigations: (1) annotations and trained YOLOv5 models from a collaborative project with professional designers developing the "Objective Portrait" method; (2) workshop protocols and student materials from an educational workshop at Design Academy Eindhoven; (3) a personal photographic dataset and LoRA models for autoethnographic inquiry; (4) ceramic datasets and LoRA models from a residency at the European Ceramic Work Centre (EKWC).<br>Materials include annotation files, trained model weights (.pt and .safetensors), workshop protocols, and analysis documentation. Due to copyright restrictions, some source images are not included; see individual chapter READMEs for details.