In India, millions of people have been protesting in recent months against new discriminatory laws and policies. And beginning in December, these demonstrations have become testing grounds for facial recognition. The Indian Express reported that police in Delhi in December used automated facial recognition technology to add to an existing database of more than 150,000 people’s faces, and taking videos to run it through their software later. Images show drones equipped with cameras flying over the heads of protesters. Once the news spread, protesters started to wear masks — usually surgical ones, made of cotton, that cover the bottom half of the face — face paint, and handkerchiefs. A cartoon circulated in response to the police action in Delhi explained “how to avoid being identified by facial recognition at protests.”
This could lead reasonable people to conclude that masks are an effective tool against an increasingly pervasive technology, both from private companies and public bodies. Large gatherings of people — football matches, concerts, shopping malls, protests — are ideal situations for law enforcement to use facial recognition technology. Consent can be assumed, rather than actively given, and the argument often used by many police forces is that training FR technology on masses of people will improve its accuracy. Police forces in the United Kingdom have been using facial recognition technology in public spaces since 2015. They’ve used it on people at Notting Hill Carnival, a massive festival in London (where the majority of attendees aren’t white), at a shopping mall in Sheffield, and during festivals in Leicester. In the US, in 2019, during a House Oversight Committee hearing, the FBI revealed that it has access to over 641 million images, from publicly available databases, which federal agencies have been using to train their own computer vision algorithms. Throwing on a mask might be an easy, effective, and immediate way to prevent such authorities from scanning your face. But that doesn’t mean it’s a credible path of resistance.
The basic way that facial recognition technology operates is by producing a mathematical reproduction of your face — your features are mapped out like a mountain or a valley. But when you wear a mask or pull up a scarf, you disfigure the topography of your face. Masks have already proven to be effective in disrupting some kinds of facial recognition technologies.
In the 2010 sci-fi novel Zero History, William Gibson described a shirt so ugly that it rendered the wearer invisible from CCTV.
As facial recognition technology becomes increasingly widespread, anti-surveillance fashion and aesthetic means of subversion have become increasingly popular too. An early tool of resistance of this kind is CV Dazzle, a kit created in 2011 by the American designer and technologist Adam Harvey, which allows people to paint their faces with colours and shapes, or fashion elaborate hairstyles, all for the purpose of warding off facial recognition. Zach Blas, another interdisciplinary artist based in London, creates physical amalgamations of people’s faces, a kind of grotesque mask that draws from many faces to make amorphous blobs. In the 2010 sci-fi novel Zero History, William Gibson described a shirt so ugly that it rendered the wearer invisible from CCTV — video surveillance, often used by businesses and increasingly, in public space — directly inspiring designer and researcher Simone C. Niquille’s “Realface Glamofage” shirts, which have close to the same function. In a bizarre twist, fans of Insane Clown Posse (Juggalos) may be put on even more FBI watchlists because Juggalo makeup — comprised of white and black blocks of color on your face — confuses computer vision algorithms.
Masks and paint are effective, low-tech and relatively easy ways to resist surveillance. But these modes of resistance are often cosmetic, Shannon Mattern, a professor at the New School who researches architecture and technology, told me. “Part of it is that it’s easily affected, an individual is capable of doing it to themselves,” she said. “It doesn’t necessarily require questioning or dismantling an entire system and it’s an issue of personal agency.”
Individual resistance becomes more difficult as surveillance technologies become more widespread. In Russia, a group of activists were arrested in mid-February for wearing anti-facial recognition dazzle makeup. A map published by the progressive advocacy group Fight for the Future shows that hundreds of airports, police departments and other institutions have begun using or announced that they are developing facial recognition programs. A few weeks ago, the Metropolitan Police in London announced that it would be rolling out live facial recognition around the city, making it the second police force in the world to do so (the first is in China). London is already viewed as the Atlantic’s greatest testing ground for facial recognition, given that over 420,000 CCTV cameras exist in the city (the next-best is Washington, D.C. which has 30,000 CCTV cameras). And this is to say nothing of the individual tools that people invite into their own homes, like Ring doorbell cameras.
“These things become a part of the world — cameras themselves become just another part of city life, and if you criticise these developments, the question then becomes, are you against security?” Simone Browne, a professor at the University of Texas at Austin, who researches the relationship between race, surveillance and aesthetics, said. “It’s seen as benign and fun and convenient. There are all those cutesy videos of people leaving cookies or something out for the mail carrier — all these emotions get tied up into how these technologies are sold.”
There are also political obstacles to mask-wearing as a tool of resistance. In some places in the U.S., like New York and Arizona, there are existing laws in place that threaten people’s rights to wear a mask as an act of protest. In China, fears over coronavirus – which has led to more and more people wearing masks in public, causing problems for both private and public uses of facial recognition — have led to private companies developing new methods of surveillance, such as thermal imaging. And for groups that are already over-surveilled, such as black and brown people who are already more likely to be misidentified by facial recognition technologies, donning a mask or a disguise is likely to attract even more attention. Black men are viewed with suspicion when they wear hoodies, and in Quebec and in France, full face coverings were banned in 2011, which stopped Muslim women from wearing niqabs and burkas if they wanted to access public services. Disguises don’t obscure social relations, and for some groups in particular, they make the disguised an object of increased scrutiny.
There are brief moments of hope, particularly when protest movements and organisations demonstrate how ways to resist surveillance come together. In Hong Kong, protesters put on a masterclass in “resisting smartness,” as Mattern puts it, throughout 2019 – they bought laser pens, which were cheap and easy to use, to shine into cameras and lines of advancing police forces, they would bring spare clothes they could change into to fool video footage of protesters, they paid for travel on public transport with one time tickets and cash, they used tinfoil to mask chips on their ID cards. But when it came down to it, the sheer number of protesters became a more effective tact of resistance than any single tool: when the government passed an emergency measure banning masks, masked crowds showed up in droves, making it impossible to identify and apprehend everyone at once.
But the prospect of crowds and crowds of people, masking up and walking around in public space, seems far off in the U.K. and the U.S., especially if it’s in direct defiance of a nationwide law. In the U.S., 59 percent of adults surveyed by the Pew Center said that they would be fine with law enforcement using facial recognition technology for assessing safety threats in public spaces, and a survey in the U.K. showed that 55 percent of adults wanted the government to restrict police use of facial recognition technology.
As I walked up to their stoops, I wondered how I would be received if I knocked on the door, only for the canvassee to see me with a mask or paint on my face.
During the December general election in the U.K., I went canvassing in a lot of neighborhoods around London. In many of the more suburban areas, I noticed more Ring doorbells than I had seen before. As I walked up to their stoops, I wondered how I would be received if I knocked on the door, only for the canvassee to see me with a mask or paint on my face (I came to the conclusion it would not end well). To be clear, Ring has stated it doesn’t yet use facial recognition — but its numerous partnerships with police departments are cause for concern. The Intercept reported last November that the company eventually plans to start deploying facial recognition technology.
It’s one thing to wear a mask when you go to a protest — or even to an area where you’re aware that you’re being watched — but it’s another thing entirely to wear one if you’re an Amazon delivery driver or jogging around a neighborhood. I don’t want to wear a mask every time I leave my house just in case there are cameras around, and I don’t think anyone else should either. I understand the impulse to take individual action against facial recognition, particularly because it seems like no one stopping its staggeringly fast rollout. The polar ice caps are melting, and shopping with a reusable grocery bag and a Keep Cup offers me some comfort as well. But even makers of tools like CV Dazzle, including Adam Harvey himself, note that the patterns have to keep evolving, as facial recognition technology becomes more attuned to the designs involved. People can keep inventing freaky masks and unflattering T-shirts, but in the tussle between ordinary people, the state, and corporations with a profit motive, even the best disguises won’t be enough.