Adversarial Fashion Garments Disrupt AI Surveillance Systems
Companies sell clothing designed to confuse facial recognition and object detection, amid growing backlash against AI-powered surveillance.
What to know
- Companies Cap_able and Urban Privacy are selling garments with adversarial patterns designed to confuse facial recognition and object detection systems.
- Patterns were developed by cybersecurity expert Bill Swearingen and tested against 11 object detection models; they work by lowering confidence scores or generating false detections.
- The trend reflects mounting public backlash against AI-powered surveillance systems citing lack of consent, unclear data storage and use, and misuse risks.
- Designers frame adversarial fashion as statement-making about surveillance rather than a reliable privacy solution.
The dispute Whether adversarial fashion represents meaningful privacy protection or primarily serves as artistic commentary on surveillance culture. · positions read across 11 posts and comments
Adversarial fashion is an important statement and symbol of resistance to mass surveillance.
-
“Privacy is a human right, and the popularity of this just goes to show that people are interested in preserving their privacy.”
Bill Swearingen · IEEE Spectrum ↗
This is a creative but ultimately limited approach—the real issue is the arms race between adversarial techniques and AI systems.
-
“AI trying to break AI is the next cybersecurity hamster wheel (which is spinning already)… this is about making a statement instead of being a serious privacy solution.”
timokissel@mastodon.world · Mastodon ↗
Bill Swearingen Cybersecurity expertRachele Didero Cap_able founder, assistant professorDaniel Preuß Urban Privacy cofounderCap_able Clothing companyUrban Privacy Fashion company
How it unfolded 1 development · click the chart to see its coverage articlesposts
-
1
IEEE Spectrum reports on commercial adversarial fashion lines and broader surveillance backlash
Coverage details Cap_able and Urban Privacy selling garments with adversarial patterns, and traces the broader movement against AI surveillance systems including ALPR mapping projects and property damage. Cap_able's dresses, pants, and tops use jacquard-woven patterns; Urban Privacy's Faception Reloaded collection uses black-and-white facial abstractions to fool detectors.
“If we're able to camouflage a person as something else, then we're obtaining our goal. We use this very visible and tangible item to talk about something that most of the time is intangible.”
— Rachele Didero -
first by HN Frontpage, 13d ago
-
T
# AI trying to break AI is the next # cybersecurity hamster wheel (which is spinning already). Also, loving this intersection of # technology and # fashion ! Funny how the machines are now dressing us, and kidding aside, this is about making a statement instead of being a serious # privacy solution. https:// spectrum.ieee.org/adversarial- fashion
2 more of the top 3 · 11 posts in this stretch
-
There is a weak component to facial recognition systems that can be exploited to make one invisible to facial recognition, but you stick out like a sore thumb to people. The exploit is to not have a human face, meaning the regular pattern of two eyes centered above a nose and mouth. The exploit is as simple as placing a sticker of a 3rd eye, a 2nd…
-
Anyone else think of "A Scanner Darkly" movie with the interpolate mask that transitions between different faces as the means for better Adversarial Fashion?Wouldn't one be able to use a hyper real mask to create false evident to frame someone? People are so blind in trusting AI that innocent people are being left in jail were primitive detective…
-
-
background
noRecognition Kickstarter project debuts at DEF CON hacker convention — Adversarial fashion project noRecognition is presented at DEF CON, showcasing patterns designed to evade surveillance cameras. Swearingen's patterns have been tested against 11 object detection models—four that search faces, two that recognize faces, and five that detect people.
-
background
Bill Swearingen develops adversarial patterns using reinforcement learning — Cybersecurity expert Swearingen creates a Python-based fuzzer targeting YOLO object detection frameworks, then develops a reinforcement learning algorithm to generate adversarial patterns that confuse facial recognition and object detection systems.
What people are saying 7 voices from 1 site · best of 11 · verbatim
- Sep 14
-
I don’t know how effective these are but it’s kinda not the point to me, the point of the clothing is a loud statement saying “I’m not okay with being surveilled all the time it’s normal and okay to to feel that way”
-
Sounds like you're describing Viola–Jones, which is from 2001. Have you tried this strategy of a third eye or similar using modern face detectors?
-
just like cvdazzle before it, this is bound to do bugger all one model into the future, and in the meantime you'd still really want to stick a pebble in one shoe and wear platforms or something to mess with gait analysisyou should read this less as convincingly adversarial fashion and more fashion inspired by the vague concept of…
-
> but you stick out like a sore thumb to peopleGood point. When I pass by the security at the airport they force me to take off my hat. Something covering my face would not be taken lightly.
-
These garments need to function in both visible and infrared environments as cameras use both. Otherwise these are not serious.
-
Face detection on the example images immediately focused on the right part and after that recognition becomes much easier when you crop out the rest. Not convinced. It’s a cool gimmick though. I like the MARPATish look of the black and white one.The most effective socially acceptable mechanism is a burqa or COVID mask or motorcycle helmet sock.
-
I hope designs like this become easier to purchase soon. It'd be fun to experiment with cameras I've got access to and I'd totally wear it out places on the weekend.Totally get that it's a cat-and-mouse thing, but I don't mind making more work for the people trying to force this on us all.