Algorithmic Apparel: Dressing for Protection in the Surveillance State

By Molly Frances Haines

Anxieties surrounding surveillance and facial recognition technology have long existed, with many concerns surrounding individual privacy, data collection, and governmental infringement on personal liberties. Surveillance technology is a booming industry, and it is projected that the market will nearly double from $114 billion in 2021 to $213 billion in 2026. Surveillance causes significant harm within communities, and in an article published by The Guardian, they cite that surveillance impairs mental health, fosters distrust between the public and the state, encourages conformity, and undermines governmental authority. As AI surveillance use increases around the globe, the concerns held about how these technologies are wielded by State and private sector powers have increased as well, especially when it comes to police use of surveillance technology.

In a piece written for Cybernews, Justinas Vainilavičius cites several reports on how police departments use sweeping surveillance networks and enormous image databases to log biometric data and track down citizens without their knowledge or consent. In the cases of the Hong Kong and Black Lives Matter protests, facial recognition technology was used to identify, track, and later punish protestors.

In the wake of these harms and concerns, an unlikely resistor has emerged: the fashion industry. Several artists and fashion companies including Cap_able, Kate Bertash, and Adam Harvey, have taken action by developing their own technologies to combat authoritarian State surveillance.

Adam Harvey is a Berlin-based artist, software engineer, and applied researcher whose work surrounds computer vision, privacy, and surveillance technologies. While his recent projects have strayed away from the fashion world, much of Harvey’s past work is grounded in anti-surveillance hair, makeup, and apparel. Starting with his CV Dazzle project, he developed looks that would obscure the wearer from the Viola-Jones face detection algorithm. At the time, CV Dazzle was the “first documented camouflage technique to successfully attack a computer vision algorithm”. Harvey expanded on this work in his later projects, HyperFace, and Stealth Wear, both of which were developed to disrupt surveillance and detection technologies. HyperFace uses “false-face” patterns to exploit facial detection algorithms while Stealth Wear aids the wearer in avoiding thermal detection technology with its silver-plated fabric. Unfortunately, Harvey’s designs were either conceptual or limited release and are no longer sold through his website.

Kate Bertash is the designer behind ‘Adversarial Fashion’ who uses patterns to inject “junk data” into Automatic License Plate Readers (ALPRs) which are used by the State and other agencies to monitor and track civilians and vehicles. The patterns were made by testing a series of modified license plate images with ALPR APIs (Application Programming Interfaces) and effectively trick ALPR readers into thinking the wearer is a car, confusing the systems. These designs can be purchased in a wide range of apparel, and Bertash also includes resources on how to make your own adversarial fashion designs.

Cap_able is an Italian company using what they call “adversarial patches” which are made by AI algorithms and woven into the designs to confuse CCTV footage. These patches trick the systems by either failing to identify their wearer, or by falsely identifying them as a giraffe, zebra, dog, teddy bear, or even donut. Cap_Able used the object detection algorithm YOLO to test their designs and while YOLO is not primarily used as facial identification software, it can be used to identify whether there is a face in the image or video. However, it can be paired with a different facial recognition algorithm that can be used to identify the found face. Cap_able’s mission is “protect the individual from the abuse of new Artificial Intelligence technologies” and “educate the population on the importance of privacy and human rights by addressing the problem of misuse of facial recognition technology”.

Fashion has a longstanding history intertwined with political activism, and what we are seeing right now with the creation of anti-surveillance fashion is part of the new wave of activism against State surveillance and violence. Although I am wishful for these fashion technologies to become the “solution” to authoritarian tech, I have some doubts about their practical applicability.

A big problem is that up until this point, many anti-surveillance fashion designs were created by artists within a niche or conceptual lens and were not made available for public purchase. Because it’s only recently that select designers began selling these kinds of garments to the public market, my concern is that it will be a long time before these styles receive mainstream wear. Currently, it is not a trend within fashion to wear anti-surveillance styles, and those who do are mostly individuals concerned with their own privacy rights. Additionally, within the mainstream, the fashion industry is driven by profits above creativity. This would be a concern if the wealthy surveillance companies and State actors were able to lobby against anti-surveillance design.

I also worry about the longevity of anti-surveillance fashion, and I think that the biggest hurdle obstructing anti-surveillance fashion is the rapid evolution of AI algorithms. Algorithms and AI are learning every day, especially within the “security” sector, and it’s only a matter of time before their detection algorithms are improved enough to bypass anti-surveillance measures. As surveillance algorithms learn, evolve, and improve, these artists and designers will have to adapt their clothing to continue tricking the systems.

Despite these apprehensions, I want to hold on to some hope. I think the most practical application for anti-surveillance fashion at this time is individual wear. However, this could very much grow into community and larger public use as the word continues to spread and designs continue to develop. I predict that more independent artists will emerge with anti-surveillance collections and designs giving the public more options for protection. I also want to hold onto my belief in the tenacity of artists in the face of State violence. The one thing surveillance lacks is a human heart and I believe that if people are passionate enough about taking a stand against authoritarian tech, then we can find a way to keep our privacy and data secure.

Image Source: Molly Haines