8 Genius Ways to Trick Surveillance Systems

The artificial intelligence startup company, Kneron, which is based in San Diego, California, has discovered a rather analog way to trick airport facial recognition technology: use a paper mask.

Facial recognition technology is becoming increasingly popular in airports as it helps decrease wait times. Think about it: a quick facial scan takes way less time than going through a traditional security checkpoint. Researchers at Kneron, however, had a sneaking suspicion that the technology wasn't as robust as it should be.

Kneron researchers donned paper masks to put their hypothesis to the test. They were able to fool systems across three different continents, including payment tablets run by the Chinese companies AliPay and WeChat, a facial recognition system at a border crossing point in China and another system in Amsterdam's Schipol Airport.

The whole thing was supervised by security guards at the airports, which gave the researchers permission to carry out these tests. Maybe don't try this one at your local airport if you enjoy not getting arrested.

Hackers designed this skirt, which is covered in license plates with parts of the Fourth Amendment (the one that deals with unreasonable searches and seizures), to trip up surveillance systems. You can snag similarly designed hoodies, T-shirts, and tanks at Adversarial Fashion's online store.

"The patterns on the goods in this shop are designed to trigger Automated License Plate Readers, injecting junk data in to the systems used by the State and its contractors to monitor and track civilians and their locations," writes Kate Rose, founder of Adversarial Fashion, on the company's website.

That, in turn, makes those plate readers less effective.

While this kind of technology can be used in positive ways, like finding missing people, recovering stolen cars, or identifying suspects in criminal cases, there are also pitfalls: It's possible for authorities to track a person in real time with nothing but their scanned license plate number. That concerns Rose, who says the U.S. Immigration and Customs Enforcement (ICE) uses the tech to find and detain undocumented immigrants.

License plate readers use infrared light to take images of a plate. Then, software examines these photos to extract and store the pertinent data. The technology can collect thousands of plate numbers per minute.

At the long-running DEFCON cybersecurity conference held in Las Vegas this year, Rose explained how her clothing creations work.

First, she tested a mind-blowing number of modified license plate images (as in, ones that don't exist yet) to find the right samples. Then, she created aesthetically pleasing fabric patterns that are read by scanning devices exactly as though they were real license plates—and cluttering the systems with crap data.

Want to create your own license plate clothing? Click here.

Since 2009, artist Adam Harvey has been working on ways to create counter-surveillance fashion, which he calls Stealth Wear.

One of his concepts (now sold out) is a scarf that you can wear like a hijab to disguise yourself from overhead drones.

Harvey makes his garments with silver-plated fabrics that can reflect thermal radiation. He believes that by 2050, countries will be able to perform full surveillance on their population for just .01% of GDP.

Harvey is also the brains behind CV Dazzle, a project that looks at how fashion can be used as a form of camouflage against facial recognition tech.

The name comes from a type of World War I naval camouflage called Dazzle. That look utilized cubist-inspired designs to "break up the visual continuity of a battleship and conceal its orientation and size," Harvey wrote in a New York Times op-ed.

CV Dazzle puts an avant-garde spin on things, with quirky and theatrical hairstyles and makeup that break up the continuity of a face—that is, it messes with the pattern of a face that an algorithm may be designed to look for while detecting people.

Usually, those algorithms are scanning for the spatial relationship between features. (For example, in most cases your eyes are usually horizontally aligned.) You can block detection, then, by creating what Harvey calls an "anti-face."

💡Try It:

1. Create asymmetry. Facial recognition algorithms are programmed to look for symmetry between the left and right sides of the face. Decrease your chances of detection by creating asymmetry, like covering your left eye with a feather or a piece of styled hair.
   
2. Use tonal inverse. Algorithms can look at the grades of skin tone and texture, which helps to locate the actual facial region. However, it relies on an assumption of what facial features look like, making it somewhat easy to trick the system. Just use hair and makeup colors that contrast with your skin tone. Apply makeup in strange directions and in unusual tones, like teal or emerald. The key is to use light colors on dark skin and vice-versa.
   
3. Cover your nose bridge. Algorithms strongly rely on the nose bridge as a key facial marker. Use accessories like gems, face paint, or creative hairstyling to conceal the area just above your nose and between your eyes.

If you really wanted to, you could walk around with a photo taped to the front of your shirt to avoid facial recognition systems from recognizing you as a person, as compared to, say, a chair.

That's based on research by Simen Thys, Wiebe Van Ranst, and Toon Goedemé at the University of KU Leuven in Belgium.

The researchers wanted to see if they could fool a human-recognizing tool called YoLo, so they created or edited various images, testing them against the artificial intelligence system until they found one that worked.

The winning image: A photo of people holding colorful umbrellas. The team rotated it and added some noise to fool the system.

As you can see in this video, when one person holds that umbrella image, the AI system doesn't detect them as a person, while another human right next to them is identified. When they switch, and the other person holds the umbrella picture, the first person is then recognized as human.

Meanwhile, the chair is still just a chair. Sorry, chair.

URME Surveillance is a company dedicated to shielding people from facial recognition technology. The idea: get as many people as possible to wear a photo-realistic, 3D-printed prosthetic "face" resembling URME founder Leo Savaggio.

When users wear the prosthetic, cameras equipped with facial recognition software identify those people as Savaggio, rather than their true identities.

Now that would be quite the flash mob.

💡Try It:

URME Personal Surveillance Identity Prosthetic: Created through state-of-the-art 3D printing, the prosthetic features a scan of the artist's face. There are even photorealistic renderings of his skin tone, texture, and hair. It sells for $200 and you can get one here.

URME Paper Surveillance Mask: Don't want to spend $200 to stick it to the man? Check out the free instructions and kit here. All you need is a printer.

New research put out in November 2019 by Northeastern University, in conjunction with the Massachusetts Institute of Technology's IBM Watson Lab in Boston, proposes a new type of T-shirt that allows people to mask themselves from people-detecting surveillance detectors by wearing tops with "adversarial" printed images on them.

As a person changes their form through movement, deforming the image, it masks them, foiling the surveillance system. The authors claim that the system can achieve up to 79 percent and 63 percent success rates in the digital and physical worlds, respectively.