This is the broader research record behind the investigation. It includes sources that directly support published findings as well as material reviewed to test claims, challenge assumptions, understand the architecture, identify contradictions, and decide what required deeper verification.
Inclusion here does not mean we accepted every statement in a source as true. Sources were evaluated according to the particular claim they were being used to examine.
Reviewed as independent government evidence concerning federal access, inadequate safeguards, and the difference between represented sharing controls and actual access.
Reviewed for natural-language video search, person search using visible characteristics, shared-camera searching, and integration of video with LPR evidence.
Reviewed for the point-in-time camera characterization and as a central source for testing component-level descriptions against network-level capability.
Reviewed for large-scale audit-log analysis, generic search reasons, and the distinction between logging activity and validating investigative purpose.
Reviewed as an advocacy investigation identifying alleged discrepancies and directing the research toward independent verification of the strongest examples.
Reviewed for the Still Image Extraction from Video workflow, which demonstrates that uploaded video can be processed to extract facial images for later searching.
Clearview AI. “Principles.”. Clearview AI. Accessed 11 Aug. 2026.
Why it was reviewed
Reviewed for law-enforcement search requirements, documentation of search purpose, audit metadata, and the uploaded-image investigative workflow.
Need the narrower record?
See only the sources cited in the published findings.
Works Cited removes the broader investigative material and keeps only the sources directly relied upon in Findings 1–12.