In the back of Artisans Asylum, a warehouse maker-space in Somerville, Mass., multimedia artist Mac Pierce presents a simple black hat, tricked out with what he calls a privacy solution to the widespread use of facial recognition technology.
Pierce has printed three angles of a face onto transfer paper, ironed the photos onto an athletic mesh and sewed the images into a seam of a traditional baseball cap.
The result is a mask-like covering that tricks facial recognition software.
“It's like looking through a couple layers of screen door,” said 26-year-old Pierce, describing the hat-wearing experience during a presentation at Artisans Asylum in January. His cap is an attempt to inspire conversation about how facial recognition technology should be used.
Watch more: The Forum Network: Mac Pierce Hacks The System
He’s not the only one finding ways to resist facial recognition systems. As applications for this kind of biometric technology develop — from Instagram filters to surveillance tracking — lawmakers and civil rights advocates are demanding privacy protections, along with clear regulations for its use by both law enforcement agencies and private businesses.
Last week, U.S. Sens. Jeff Merkley, D-Ore., and Corey Booker, D-N.J., proposed the Ethical Use of Facial Recognition Act, which would pause the government’s use and funding of facial recognition technology until “appropriate guidelines and limitations” are passed.
In Massachusetts, House Democrats introduced a similar bill on Beacon Hill in January 2019, but its scope goes beyond facial recognition to include other types of biometrics, such as iris scanning and gait tracking. The bill would place a moratorium on law enforcement’s ability to “acquire, possess, access, or use any biometric surveillance system.”
But neither bill would regulate commercial use of the tech, meaning businesses could still use biometric tools to identify or track customers.
Meanwhile, DIY-style solutions have popped up around the globe. Artist Adam Harvey designed a make-up schema called CV Dazzle that uses hair placement and glitter to block the technology; LED glasses, created by professors at Tokyo's National Institute of Informatics, thwart face mapping with infrared lights; and products like URME prosthetic masks function similarly to Pierce’s cap by using 3D prints to fool the software into reading another face.
Pierce has dubbed his creation the “ Opt-Out Cap” to emphasize the lack of consent around the use of facial recognition.
“The individual is automatically opted in without being given the choice,” said Pierce, “Like for emails … you get a little unsubscribe button. With most of these facial recognition technologies, there is no unsubscribe form.”
Pierce said there have been over 100,000 visits to the Opt-Out Cap’s assembly page, but said he’s never spotted the cap being used in real life.
Although the hat successfully conceals one's identity from recognition software, Pierce explained that any government or corporation with enough resources could find another way, like gait tracking or reading heart beat data, to ID individuals. The Opt-Out Cap was ultimately designed to raise questions about the technology, he said.
“Even if this isn't a magic bullet to protect you in, like, a facial recognition dystopia scenario,” he said, “it gets you thinking about the technology overall.”
Facial recognition technology is now commonly used to secure our phones, swiftly recognize friends in photos on social media, identify suspects in criminal investigations and function as a security measure in airports. It has even been used to track attendance in schools.
According to a report by Modor Intelligence published in January, the facial recognition market was valued at $5.07 billion last year and is expected to double by 2025.
Although there are convenience and security benefits, studies have raised concerns about the inaccuracy of the technology. In 2018, both the American Civil Liberties Union and MIT Media Lab released data about the technology’s limitations, revealing high rates of false positives and misidentifications — particularly of women, children and people of color.
While legislation is pending at the state and federal levels, municipalities in Massachusetts have already begun responding to the fast-growing applications of face recognition.
Somerville banned the governmental use of the technology in June 2019, the second city to do so in the country. Over the past year, similar bans have been passed in Brookline, Northampton and Cambridge, all citing privacy concerns and a right to anonymity.
“It’s important that folks understand where we are now,” said Somerville City Councilor Ben Ewen-Campen, who proposed the city’s ban. “There are absolutely no regulations about how this technology can be used.”
He likened facial recognition to a bar code that each citizen wears, able to be tracked.
But some see these bans as missing the crux of the problem. Dr. Tanya Zlateva, dean of Boston University’s Metropolitan College and co-founder of the Center for Reliable Information Systems and Cyber Security, told WGBH News that the most serious problem is really how easily accessible this personal data is and that it can be used without consent.
Facial recognition technology is cropping up so quickly compared to other types of biometrics because it's relatively easy and cheap to construct, she said. But just banning facial recognition is a symptom of the public misunderstanding how the technology functions, she said, because it doesn't address the fact that the data is still out there.
Zlateva said she believes the solution lies not in banning facial recognition or any one type of biometric, but in regulating how we consent to the sharing and tracking of identifying data — the building blocks of these technologies.
“When we talk about privacy and security, it’s not just for face recognition or iris recognition or for date of birth and social security,” she said. “It’s how we deal with protecting data that the person to whom it belongs has not explicitly given permission to use.”
Facial recognition is built from three main components: First, hardware — like a camera — views a person’s face; second, a database provides a pool of faces to match data to; and third, software runs algorithms that read the face, scan the database and accomplish a task, such as identifying or authenticating.
Municipalities banning the technology “is not going to work,” Zlateva said, because it won't stem the root problem of that data being widely available.
While specific regulations around the use of biometric technology and personal data seem far off, Pierce offers his creation as a last resort for the video cameras pointing at public streets and hanging in retail stores. An exhibit of the Opt-Out Cap will be on display at Artisans Asylum until mid-March.
“I wanted to put together something that could give the individual the option to disengage with those systems — of opting out,” said Pierce.
