AI-Generated Pattern Evades Flock and Clearview Detection

Flock Safety’s automated license-plate readers and Clearview AI’s facial-search tools have become part of a widening U.S. surveillance infrastructure, turning camera footage into searchable data for law enforcement. The spread has intensified concerns over tracking without consent, false matches and weak oversight. Cybersecurity researcher Bill Swearingen’s noRecognition project applies adversarial machine learning to generate patterns for clothing or vehicle surfaces that disrupt the object-detection step before a system can identify and catalog a person or car.

At DEF CON in Las Vegas on Aug. 7, 2026, Swearingen covered a 2009 Toyota Yaris with an AI-generated pattern and drove it past a Flock camera, preventing the software from flagging it as a vehicle. He said the yearlong project has run about 31 million tests, now produces a new pattern every minute, and has defeated 11 open-source detection models associated with Flock, Axon and Clearview AI. The camera still records video; the pattern targets automated detection and indexing, not the lens itself.