Clothing 'Wearable Algorithm' Makes CCTVs Identify Humans as Animals, Confuses AI Recognition

Italian startup Cap-able has come up with its Manifesto Collection, which includes a £252 shirt, a £370 sweater, and a pair of £245 jogging pants. What makes this specific collection stand out, however, is that it is a wearable algorithm that can help protect one's identity.

Privacy Solution

Metro reports that the fabric is created with adversarial patches that are capable of protecting facial data. More than this, it can also trick AI into identifying human wearers as giraffes, dogs, or zebras.

According to the New York Post, CEO Rachele Didero expresses on the company's official website that in a world where data has become the "new oil," Cap-able caters to privacy concerns and leads to discussions regarding how individuals should be protected from biometric recognition camera misuse.

Didero states that the concern is present throughout one's daily life and that it is something that involves individuals from around the world. If the privacy concern is neglected, individuals' rights regarding expression, association, and movement could be frozen.

Cap-able also notes that their Manifesto Collection is the first collection that provides knitted clothing that serves as a facial recognition shield.

Adversarial patches are only printed until now. Cap-able, however, was able to patent a process to have the algorithm included and woven into the knitted fabric.

Wearable Algorithm Tricks Artificial Intelligence Into Thinking That the Wearer Is an Animal

According to Fox 10 Phoenix, Rachel Didero had the design idea as she took up her Master's in the Fashion Institute of Technology. There, she read a narrative regarding how certain tenants in Brooklyn argued with their landlord, who planned to set up a facial recognition device at the building entrance. She was then inspired to come up with something that could grant people a choice.

Cap-able combined both fashion and engineering for its Manifesto Collection. The company comes up with an algorithm and evaluates each image through YOLO, which is an open-source system for object detection that makes use of neural networks to speedily and accurately read objects, to see if their algorithm could outsmart systems for facial recognition. They could create a physical pattern with the Computerized Knitwear Machine and make the final output with Egyptian cotton.

Though the designs could be alluring, they are quite expensive. Fox 10 Phoenix reports that a short-sleeve crew-neck shirt would cost almost $300.

Editor of Our Connect, Lucy Hedges, expresses that Cap-able may help people feel that their personal data is being protected. Hedges notes that if the company gets it right, their innovation may be revolutionary. However, if they fail, all that people get is ugly and expensive knitwear.

Check out more news and information on Artificial Intelligence in Science Times.