Cybersecurity researcher Bill Swearingen says he has developed computer-generated patterns designed to interfere with AI-powered surveillance systems, including some person detectors and license plate readers. His noRecognition project uses reinforcement learning to generate patterns that can reduce or prevent automated detection without stopping cameras from recording footage. Swearingen spent about a year testing the approach, running roughly 31 million experiments against multiple detection systems. The systems tested include software associated with surveillance products used for license plate detection, body-worn cameras and facial recognition. Swearingen is also seeking funding to produce clothing carrying the patterns, including T-shirts and hoodies, with vehicle wraps potentially following later.