In May 2019, an agent at the Department of Homeland Security received a
trove of unsettling images. Found by Yahoo in a Syrian user’s account,
the photos seemed to document the sexual abuse of a young girl. One
showed a man with his head reclined on a pillow, gazing directly at the
camera. The man appeared to be white, with brown hair and a goatee, but
it was hard to really make him out; the photo was grainy, the angle a
bit oblique. The agent sent the man’s face to child-crime investigators
around the country in the hope that someone might recognize him.
When an investigator in New York saw the request, she ran the face
through an unusual new facial-recognition app she had just started
using, called Clearview AI. The team behind it had scraped the public
web — social media, employment sites, YouTube, Venmo — to create a
database with three billion images of people, along with links to the
webpages from which the photos had come. This dwarfed the databases of
other such products for law enforcement, which drew only on official
photography like mug shots, driver’s licenses and passport pictures;
with Clearview, it was effortless to go from a face to a Facebook
account.
The app turned up an odd hit: an Instagram photo of a heavily muscled
Asian man and a female fitness model, posing on a red carpet at a
bodybuilding expo in Las Vegas. The suspect was neither Asian nor a
woman. But upon closer inspection, you could see a white man in the
background, at the edge of the photo’s frame, standing behind the
counter of a booth for a workout-supplements company. That was the
match. On Instagram, his face would appear about half as big as your
fingernail. The federal agent was astounded.
The agent contacted the supplements company and obtained the booth
worker’s name: Andres Rafael Viola, who turned out to be an Argentine
citizen living in Las Vegas. Another investigator found Viola’s Facebook
account. His profile was public; browsing it, the investigator found
photos of a room that matched one from the images, as well as pictures
of the victim, a 7-year-old. Law-enforcement officers arrested Viola in
June 2019. He later pleaded guilty to sexually assaulting a child and
producing images of the abuse and was sentenced to 35 years in prison. (Viola’s lawyer did not respond to multiple requests for comment.)
At the time, the use of Clearview in Viola’s case was not made public; I
learned about it recently, through court documents, interviews with
law-enforcement officials and a promotional PowerPoint presentation that Clearview made.
The case represented the technology’s first use on a child-exploitation
case by Homeland Security Investigations, or H.S.I., which is the
investigative arm of Immigrations and Customs Enforcement. (Such crimes
fall under the agency because, pre-internet, so much abuse material was
being sent by mail internationally.) “It was an interesting first foray
into our Clearview experience,” said Erin Burke, chief of H.S.I.’s Child
Exploitation Investigations Unit. “There was no way we would have found
that guy.”
Few outside law enforcement knew of Clearview’s
existence back then. That was by design: The government often avoids
tipping off would-be criminals to cutting-edge investigative techniques,
and Clearview’s founders worried about the reaction to their product.
Helping to catch sex abusers was clearly a worthy cause, but the
company’s method of doing so — hoovering up the personal photos of
millions of Americans — was unprecedented and shocking. Indeed, when the
public found out about Clearview last year, in a New York Times article I wrote, an immense backlash ensued.
Facebook, LinkedIn, Venmo and Google sent cease-and-desist letters to
the company, accusing it of violating their terms of service and
demanding, to no avail, that it stop using their photos. BuzzFeed
published a leaked list of Clearview users,
which included not just law enforcement but major private organizations
including Bank of America and the N.B.A. (Each says it only tested the
technology and was never a client.) I discovered that the company had
made the app available to investors, potential investors and business
partners, including a billionaire who used it to identify his daughter’s date when the couple unexpectedly walked into a restaurant where he was dining.
Computers once performed facial recognition rather imprecisely, by
identifying people’s facial features and measuring the distances among
them — a crude method that did not reliably result in matches. But
recently, the technology has improved significantly, because of advances
in artificial intelligence. A.I. software can analyze countless photos
of people’s faces and learn to make impressive predictions about which
images are of the same person; the more faces it inspects, the better it
gets. Clearview is deploying this approach using billions of photos
from the public internet. By testing legal and ethical limits around the
collection and use of those images, it has become the front-runner in
the field.
After Clearview’s activities came to light, Senator Ed Markey of
Massachusetts wrote to the company asking that it reveal its
law-enforcement customers and give Americans a way to delete themselves from Clearview’s database.
Officials in Canada, Britain, Australia and the European Union
investigated the company. There were bans on police use of facial
recognition in parts of the United States, including Boston and
Minneapolis, and state legislatures imposed restrictions on it, with
Washington and Massachusetts declaring that a judge must sign off before
the police run a search.
In Illinois and Texas, companies already had to obtain consent from
residents to use their “faceprint,” the unique pattern of their face,
and after the Clearview revelations, Senators Bernie Sanders and Jeff
Merkley proposed a version of Illinois’s law for the whole country.
California has a privacy law giving citizens control over how their data
is used, and some of the state’s residents invoked that provision to
get Clearview to stop using their photos. (In March, California
activists filed a lawsuit in state court.) Perhaps most significant, 10
class-action complaints were filed against Clearview around the United
States for invasion of privacy, along with lawsuits from the A.C.L.U.
and Vermont’s attorney general. “This is a company that got way out over
its skis in an attempt to be the first with this business model,”
Nathan Freed Wessler, one of the A.C.L.U. lawyers who filed the
organization’s lawsuit, in Illinois state court, told me.
It seemed entirely possible that Clearview AI would be sued, legislated or shamed out of existence. But that didn’t happen....
