This $88 Hawaiian Shirt Blinds AI Surveillance Cameras Completely

German artist Simon Weckert designed a button-down garment covered in chaotic pink, green, and orange patterns specifically to confuse machine vision systems. The outfit was born as a direct response to a pilot project deploying automated AI video surveillance at Kottbusser Tor square in Berlin.

The fabric targets the very first stage of machine perception where software attempts to draw a bounding box around a human silhouette before running facial or gait recognition. Sharp colors overload these perceptual layers while overlapping shapes fracture body contours, preventing the program from connecting arms, torso, and head into a single recognizable figure.

Weckert utilized the open-source YOLO model to build the textile and successfully tested it in June 2026 right in front of an active surveillance camera, keeping total production costs down to just $88. This follows a similar recent trend where cybersecurity researcher Bill Swearingen created adversarial patterns on film to protect cars from tracking algorithms, successfully tricking eleven open-source detection models as well as Flock license plate readers and Axon police body cameras.

Garish holiday clothing is quietly transforming from a fashion crime into a vital civic utility. Humanity always finds a way to weaponize bad taste against authoritarian convenience.

Source: Dexerto