An AI-generated car wrap kept Flock cameras from classifying a Yaris

A 2009 Toyota Yaris in an AI-generated wrap drove past a Flock camera, and the software could not work out what it was looking at.

The experiment

Cybersecurity researcher Bill Swearingen built a project called noRecognition: patterns generated specifically to confuse the computer vision behind surveillance systems. For the public test he teamed up with Donut Media, covered the Yaris in the pattern and drove it past a Flock camera at DEF CON in Las Vegas. The camera still records the car. The software just struggles to identify what it is seeing.

Swearingen says the patterns defeated 11 open-source surveillance algorithms, and reportedly also confuse Axon body cams and Clearview AI. Donut Media’s video of the experiment is out now.

What it does not do

This is not an invisibility cloak. Results vary with the camera, the angle, the light and the software, and a human reviewing the footage sees a very loud car immediately. What the test proves is narrower: the automated layer, the one that scans every passing vehicle without a human in the loop, can be broken by paint.

Why it matters

Flock’s network has been at the center of a growing backlash, from misuse of its search tools to cameras being physically cut down. A legal, reversible, street-tested way to degrade automated recognition adds a new front: not destroying the cameras, just declining to be machine-readable. An arms race between patterns and detection models is the predictable next step.