A secretive startup that promotes a massive international universal facial recognition database seeded from more than three billion images is facing pushback from tech firms as it tries to woo more law enforcement agencies.
The company, called Clearview AI, went from near-complete obscurity to national headlines following a report published by the New York Times in January. The Times described Clearview as a “groundbreaking” facial recognition service. A user imports a photo of a person, and then the app shows “public photos of that person, along with links to where those photos appeared,” the NYT explains.
Clearview claims to have agreements with 600 law enforcement agencies for use of its services, and the company says it has a set of three billion public photos to match new images against. Those photos come from a wide array of sources, and the sources are ticked off.
Twitter was the first major tech firm to tell Clearview to quit it. A few days after the NYT first reported on the company, Twitter sent a letter to Clearview saying that its actions were in violation of Twitter policy. The letter included demands that Clearview stop scraping images or other data from the platform “for any reason” and delete data it had already collected.
Where Twitter went, others rapidly followed. In the days since, Google told Clearview to stop scraping YouTube, Facebook told it to stop gathering data from Facebook or Instagram, and Microsoft told it to stop scraping LinkedIn. Peer-to-peer payment service Venmo also told Clearview to stop.
When the New York Times first published its profile on Clearview, company founder and CEO Hoan Ton-That was evasive. Reporter Kashmir Hill wrote:
When I began looking into the company in November, its website was a bare page showing a nonexistent Manhattan address as its place of business. The company’s one employee listed on LinkedIn, a sales manager named “John Good,” turned out to be Mr. Ton-That, using a fake name. For a month, people affiliated with the company would not return my emails or phone calls.
Clearview, however, was clearly paying attention. When Hill asked several police officers to run a photo of her through the app, the company began contacting those law enforcement agencies asking if they had been talking to the media—apparently Clearview does, indeed, know a reporter when it sees one.
The company began marketing its services to law enforcement agencies in 2017 and 2018, at which point it attracted investors. Among the first to invest was infamous venture capitalist Peter Thiel—best known for his involvements with PayPal, Facebook, and Palantir, and for financing the lawsuit that eventually resulted in the bankruptcy of media company Gawker. As with Ring, Amazon’s line of home surveillance products, partnering with police seems to have been just the ticket for Clearview.
Though the company told the Times that more than 600 law enforcement agencies now use its services, it did not provide a list. “Federal and state law enforcement officers said that while they had only limited knowledge of how Clearview works and who is behind it,” the NYT wrote, “they had used its app to help solve shoplifting, identity theft, credit card fraud, murder and child sexual exploitation cases.”
The Times more recently ran another report on Clearview’s role in those child exploitation cases. One investigator in Canada told the paper that Clearview represented “the biggest breakthrough in the last decade” for law enforcement trying to track down those who exploit children. Privacy advocates, however, expressed concern to the NYT about the fact that Clearview stores images—potentially extremely explicit and sensitive images—investigators upload to the platform. Clearview’s security practices are as yet unproven, and those images are “retained forever” by default, though users can change their settings so the images are only held for 30 days.
In an interview with CBS, Ton-That argued that his firm’s work is protected by the First Amendment and also that Clearview doesn’t do anything Google doesn’t.
“The way we have built our system is to only take publicly available information and index it that way,” he said.
Ton-That added, “Google can pull in information from all different websites… So if it’s public and it’s out there and could be inside Google search engine, it can be inside ours as well.”
Meanwhile, some of the company’s claims about its partnership with law enforcement seem unproven. Clearview boasts that its technology helped lead to the arrest of a would-be terrorist in New York City, for example. But when BuzzFeed reporters contacted the NYPD to ask about the case, the police said Clearview had nothing to do with it.
“The NYPD did not use Clearview technology to identify the suspect in the August 16th rice cooker incident,” the agency told BuzzFeed. “The NYPD identified the suspect using the Department’s facial recognition practice where a still image from a surveillance video was compared to a pool of lawfully possessed arrest photos.” The New York police also denied to BuzzFeed that Clearview played a role in other cases the company referenced in its marketing materials.
As one might expect, lawmakers and privacy advocates are pushing back heavily on Clearview.
Sen. Edward Markey (D-Mass.) sent the firm a list of questions (PDF) about its platform. In addition to seeking a list of “all law enforcement or intelligence agencies” that use or have expressed interest in using Clearview, Markey asked the company to explain how it tests its software, how it secures its data, and if it will commit to removing individuals who ask to be scrubbed from its database.
“The ways in which this technology could be weaponized are vast and disturbing,” Markey wrote. “Widespread use of your technology could facilitate dangerous behavior and could effectively destroy individuals’ ability to go about their daily lives anonymously.”
The office of Sen. Ron Wyden (D-Ore.) set up a meeting with Ton-That to discuss Clearview’s privacy implications. Wyden said in a series of tweets that he found the NYT’s description of Clearview “extremely troubling.”
“This story reads like one of the more disturbing episodes of Black Mirror,” Wyden said. “Americans have a right to know whether their personal photos are secretly being sucked into a private facial recognition database.”
The office of New Jersey Attorney General Gurbir Grewal issued a cease and desist order to Clearview, barring it from advertising “in any manner that suggests an affiliation” with state officials. “Your actions in this regard are unauthorized and appear to be in violation of numerous New Jersey and federal laws,” counsel Christopher Edwards wrote.
The company is also facing a lawsuit in Illinois alleging violation of that state’s Biometric Information Privacy Act, which requires entities that collect and use state residents’ biometric information to provide notice and gather consent from those individuals.
Clearviews offers “a massive surveillance state with files on almost every citizen, despite the presumption of innocence,” the complaint (PDF) says. “Clearview’s technology poses a grave threat to civil liberties.”
A coalition of 40 privacy and civil rights organizations, headed by the Electronic Privacy Information Center (EPIC), sent a letter (PDF) to several members of Congress asking for their intervention.
The letter described several known issues with bias and accuracy in facial recognition systems before adding, “While we do not believe that that improved accuracy of facial recognition would justify further deployment, we do believe that the obvious problems with bias and discrimination in the systems that are currently in use is an additional reason to recommend a blanket moratorium.”