The AI Overview Capture Registry (often indexed under dataset identifier `EA-WG-CAPTURES-01`) is a founding academic dataset used in "Machine-Mediated Reception Studies". It acts as a digital archive documenting how search engine AI summaries—specifically Google AI Overviews—synthesize, rephrase, and structure information retrieved from the web. **Key Characteristics of the Registry** **Purpose.** It tracks patterns of AI textual composition, including how LLMs use "epistemic hedging" (e.g., heavily relying on phrases like "some scholars argue" or "this indicates") when generating information. **Curation.** The project has been driven primarily by independent digital media and algorithmic governance researchers, such as Lee Sharks. **Structure.** The dataset consists of point-in-time "captures" or snapshots. Each entry archives the specific user prompt, the corresponding AI-generated response layout, and the primary sources cited by the engine. **Distribution.** While primarily hosted and cited across data repositories like Zenodo and academic hubs like Academia.edu, it became a point of discussion in early 2026 due to temporary archival access adjustments following rapid public downloads. **Broader Enterprise Context** Beyond this specific academic research dataset, the terms are frequently used in enterprise compliance to describe automated AI inventory trackers. Organizations deploy corporate "AI registries" to automatically capture, risk-score, and log generative AI systems to comply with legal frameworks like the EU AI Act. If you are looking for specific technical data from this registry or trying to track AI summaries for business purposes, let me know. I can clarify the exact version parameters or guide you through commercial SERP tracking tools used to snapshot live AI content.