How to Block Surveillance Cameras with Adversarial Patterns – Protect Your Privacy!

Our Take

Bilal Khan's video does a solid job breaking down Bill Swearingen's 'noRecognition' project, which uses computer-generated adversarial patterns to confuse facial recognition algorithms. It's a clever piece of applied research and worth a watch if you're curious about how machine learning models can be tricked by patterns humans barely register. Credit where it's due: this is the kind of grassroots privacy engineering we like to see more of.

But here's the catch for our readers: Flock Safety's ALPR network isn't reading faces. It's reading plates, vehicle make, color, bumper stickers, even dents — building a searchable log of where your car has been, regardless of who's driving. An adversarial t-shirt won't do a thing to stop a camera bolted to a pole that's cataloging your plate hundreds of times a month. The threat model is different, and that distinction matters because it shapes what 'protection' actually looks like.

Research like Swearingen's matters as a proof of concept that algorithmic surveillance isn't infallible — but the real fight against ALPR networks is political and legal, not just technical. Check our camera map to see what's logging your plate in your own neighborhood, and head to our take-action page if you want to push back on the policies that let this surveillance infrastructure spread unchecked.