AI Overview Capture Registry v8.3

The AI Overview Capture Registry is a primary empirical dataset documenting how Google's AI composition layer — AI Overview and AI Mode — receives, compresses, and redistributes scholarly content. Version 8.3 contains 176 captures recorded between June 7 and June 18, 2026, encompassing 205 screenshot images across two gallery mirrors. Each capture records a query, the AI-generated response surface, a match-type assessment (exact match, partial, compositional bystanding, or zero composition eligibility), and a descriptive analysis of how the composition layer treated the source material. The dataset is a foundational instrument of Machine-Mediated Reception Studies, demonstrating the field's core method: systematic empirical observation of machine composition behavior under documented conditions. The registry constitutes primary research data in which AI-generated content is the object under study, not the research output — the AI systems produced the responses being documented; the research design, capture methodology, classification schema, and analysis were human-governed. At the time of its removal from Zenodo, the Capture Registry had received over 1,000 downloads, making it the most-accessed work in the Crimson Hexagonal Archive. The concept of 'compositional bystanding' — where a source holds top organic ranking but receives zero composition eligibility in the AI layer — was documented through this dataset as a new failure mode in scholarly visibility.