As It Happens

London artists don makeup to thwart facial-recognition technology

Once a month, a brigade of artists marches through the heavily surveilled streets of London with colourful geometric shapes painted on their faces.

The Dazzle Club describes the colourful geometric shapes as 'reverse contouring'

Emily Roderick is an activist with the Dazzle Club, a group of London activists who are wearing anti-facial recognition make-up in defiance of Metropolitan police. (Cocoa Laney)
Listen5:59

Transcript

Once a month, a brigade of artists marches through the heavily surveilled streets of London with colourful geometric shapes painted on their faces.

They're called the Dazzle Club and they're part of an open-source project that tests methods of thwarting facial recognition technology which is being deployed on the streets of England's largest city by the Metropolitan police. 

"You paint your face in a specific way ... that will then scramble your face against automatic facial recognition systems so that your face is no longer recognized," Dazzle Club co-founder Emily Roderick told As It Happens guest host Helen Mann. 

"We describe it as reverse contouring, or kind of you're really looking to de-emphasize parts of your face that are normally quite prominent."

Law enforcement employing the technology 

Facial recognition software works by mapping out facial features and comparing them to existing databases to find matches. 

The technology has grown in reach and scope in recent years, moving from public apps that tag photos, to massive networks used by law enforcement agencies.

Last week, London's police force announced it will be using what's called "live facial recognition" cameras as part of the city's already widespread CCTV surveillance network, drawing the ire of British privacy advocates. 

"As a modern police force, I believe that we have a duty to use new technologies to keep people safe in London. Independent research has shown that the public support us in this regard," assistant commissioner Nick Ephgrave told the Guardian.

The newly-formed group of artists in London are protesting the Metropolitan police's use of facial recognition technology by painting colourful geometric shapes on their faces. (Megan Jacob)

In Canada, the RCMP uses Clearview AI's facial recognition software, which can match images against a massive database of billions of photos scrubbed from public websites like Facebook and Instagram.

The powerful software, which the New York Times has reported extensively on, can unearth a person's name, phone number, address or occupation based on a single picture. 

The Mounties issued a statement earlier this month confirming it had been using Clearview AI for at least the previous four months — weeks after denying it to the CBC.  It has since said it will only use it  "in very limited and specific circumstances."

Accuracy questioned 

Roderick says one of her biggest concerns is that the technology in London will be used to falsely identify suspects, especially people of colour. 

The Met maintains its system is 70 per cent effective at spotting suspects, but an independent review by Essex University surveillance expert Pete Fussey found it was accurate in just 19 per cent of cases.

Meanwhile, an analysis of three unnamed commercial facial recognition tools by MIT's Media Lab in 2018 found the software to be 99 per cent accurate at identifying light-skinned men — but made mistakes in 21 per cent and 35 per cent of cases for dark-skinned women.

"We are against the idea of use of the usage of facial recognition purely because of the technology not being advanced enough at this point," Roderick said. 

"There's a real big issue that we're, you know, deploying these systems within our streets when really they're not fully functioning —  if that's something that society actually wants. I mean, ideally, it'd be great if we didn't have our facial recognition at all."

Accuracy aside, she says the group is also interested in exploring the idea of surveillance more broadly. 

"Primarily, the Dazzle Club is there to talk about and discuss the use of facial recognition. But I think also we're there to talk about the use of surveillance within the city more widely," Roderick said.

"It's opening up a really important discussion about the city. And we've had a lot of interest over the past few months now. "

Facial recognition resistance

The Dazzle Club's method was spearheaded by artist Adam Harvey, who coined the term "computer vision dazzle." 

It's one of several methods developed by artists around the world to thwart facial recognition technology, including Jing-Cai Lu's wearable face projector, which beams another face over the wearer's, or Jip van Leeuwenstein's clear plastic mask that distorts a user's face with ridges.

Roderick says the Dazzle Group has tested their makeup's effectiveness using smartphone and laptop software that uses facial recognition — like Facebook, Instagram or face-scanning security apps.

"I know it's not necessarily testing on exactly the surveillance system that might be used within London streets, but we're getting close to that, and we can still be able to see when the face paint works and when it doesn't," she said. 

Still, she admits facial recognition technology could eventually adapt to the dazzle method. 

"It's this strange arms race in a way," she said.


Written by Sheena Goodyear with files from CBC News. Interview produced by Jeanne Armstrong.

Comments

To encourage thoughtful and respectful conversations, first and last names will appear with each submission to CBC/Radio-Canada's online communities (except in children and youth-oriented communities). Pseudonyms will no longer be permitted.

By submitting a comment, you accept that CBC has the right to reproduce and publish that comment in whole or in part, in any manner CBC chooses. Please note that CBC does not endorse the opinions expressed in comments. Comments on this story are moderated according to our Submission Guidelines. Comments are welcome while open. We reserve the right to close comments at any time.