As artificial intelligence (AI) continues to be used widely for surveillance in public places, concerns are rising about increasing intrusions into personal privacy. As a response, an Italian startup called Cap_able has created a line of anti-facial recognition clothing designed to protect wearers from AI software without having to cover their faces. Its unique designs are based on an “adversarial patch,” a pattern developed by AI algorithms to confuse facial recognition software. The patches are designed to resemble animals like giraffes and zebras since most cameras are built to filter out information about animals.
Facial recognition technology is used globally by governments and corporations to track and identify individuals. There is perceived value in capturing criminals efficiently, however, worries persist about privacy issues for ordinary citizens. For example, in Australia, banks, phone companies, immigration authorities, and law enforcement agencies use facial recognition AI for purposes such as identity verification. In Adelaide, South Australian Police have been using an AI tool, the “NEC NeoFace system,” for facial recognition since 2019.
The designs on Cap_able’s clothing reportedly work 60 to 90% of the time when tested against commonly-used facial recognition algorithms. The concept originated from the Ph.D. research of Cap_able co-founder Rachele Didero at the University of Milan. The clothes interrupt the facial recognition software, albeit temporarily, as companies and governments could eventually classify the clothing as human in their databases.
However, not everyone may be able to afford this apparel as a method of safeguarding their privacy. With a significant price tag of €420 (AU$675) for a jumper, these garments are currently only within the reach of wealthier individuals. Consequently, the high cost of dissent in today’s society is a concern. However, as surveillance increases, the cost of these garments may eventually be seen as a small price to pay for the protection of personal privacy. In summary, while the debate around privacy and surveillance continues, methods such as anti-facial recognition clothing are emerging as potential ways to retain anonymity in an AI-controlled environment.
