Clothing Brand noRecognition Scrambles AI Security Cameras

A new clothing brand named noRecognition has emerged, offering garments designed to confuse and scramble AI security cameras. This initiative is a direct response to the increasing prevalence of public surveillance systems, aiming to provide a form of adversarial fashion that baffles the sensors of tools like Axon body cams and Flock cameras. The patterns employed by noRecognition are engineered to trick AI sensors into misidentifying them as animals, common objects, or reflective fabric, thereby disrupting their surveillance capabilities. The brand promises that its patterns are effective enough to scramble at least 11 different computer vision models.

Bill Swearingen, the founder of noRecognition and a former chief intelligence officer, developed the concept after conducting extensive digital testing. He ran 31.7 million tests to identify patterns that could simultaneously confuse multiple AI models, with the ultimate goal of creating a universal adversarial camouflage. Swearingen utilized artificial intelligence extensively throughout this process, including building and training his own AI model. This model was then employed to generate and test various adversarial pattern geometries, ultimately determining the most effective designs. Swearingen highlighted the complexity of this challenge, noting that a pattern effective against one AI model often fails against others. He stated, "I have beat every model I have tested, so beating a single model is a solved problem for me. The search now is finding that one pattern that works across many models at once."

noRecognition was publicly introduced at the Def Con cybersecurity conference in Las Vegas on August 7. Concurrently, Swearingen launched a Kickstarter campaign with an initial fundraising goal of $5,000. The campaign significantly exceeded expectations, raising over $40,000. Swearingen expressed his gratitude for the overwhelming support, describing the response as "humbling." He confirmed that the funds raised will be allocated towards acquiring fabric, cameras, and compute resources necessary for further development and production.

The initial limited-edition clothing line from noRecognition includes a versatile buff that can be worn as a neck gaiter, a T-shirt, and a sweatshirt. The collection also features stickers and patches. Each item in the line is part of a 50-item production run, with each garment featuring a unique pattern generated specifically for an individual. Swearingen's website showcases examples of the sophisticated patterns developed by the noRecognition model, illustrating the intricate designs intended to disrupt AI-powered surveillance.