2027 Fashion Prediction: Adversarial

Adversarial is the new black. Brought to you by cybersecurity researcher turned fashionista Bill Swearingen.

His project, noRecognition, creates computer-generated patterns that confuse AI-powered surveillance cameras and license plate readers. Put one of his patterns on a person or vehicle, and the system may look right at it and see nothing worth reporting.

After 31 million tests, Swearingen’s reinforcement-learning model beat 11 open-source detection systems, including systems used in Flock license plate readers, Axon body cameras, and Clearview AI. At Def Con he made a vehicle effectively disappear from a Flock camera.

To us, the patterns look like chaotic geometric art. To a surveillance system, they create a sort of invisibility cloak. Which is why this could become de rigueur among privacy advocates, protesters, hackers, and the fashionably surveillance-averse.

Swearingen sees privacy as a fundamental right and the pattern as a way to opt out of being tracked everywhere we go. He’s already crowdfunding T-shirts and hoodies at Kickstarter, with vehicle wraps likely in the future. He’s also keeping his most effective patterns offline so camera companies can’t immediately train their systems to recognize them.

And here's where it gets really fun. The model keeps creating new and more effective patterns. The surveillance systems learn. It’s an arms race. With hoodies.

The next big fashion statement won't be a logo, color, or celebrity collab. We think it's adversarial prints that let you stroll past a street camera without adding to a database.

So yes, adversarial is the new black. And soon, you might see it everywhere - even if the cameras don’t. 😉