[REVISION HISTORY]

AI vehicle surveillance evasion research

Updated 1 time since CLSTR started tracking revisions of this situation.

What changed

2026-08-22 23:36 UTC → 2026-08-26 12:51 UTC · added ~~removed~~

Cybersecurity expert Bill Swearingen has developed a method called ‘noRecognition’ to evade automated vehicle detection and license plate reader (ALPR) systems. By applying computer-generated geometric patterns to a vehicle, the software fails to recognize the presence of a vehicle during the initial detection phase, even though the vehicle remains clearly visible to human observers. Swearingen, a co-founder of the SecKC security meetup, tested over 31 million patterns using a fuzzer to identify effective designs. During a demonstration at the Def Con conference in Las Vegas, a Toyota Yaris equipped with these patterns successfully passed a Flock camera without being logged by the detection software. The project was motivated by concerns regarding unjust arrests linked to surveillance technology and highlights vulnerabilities in AI-based urban monitoring architectures. The ‘noRecognition’ project utilizes these physical adversarial patterns—abstract designs intended to disrupt automated vision models—to make cars ‘invisible’ to AI-powered surveillance. Notably, the technology does not obscure license plates, which remain visible and compliant with regulations; instead, the pattern is designed to prevent the algorithm from identifying the object as a vehicle in the first place.

Versions

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.