Clearview AI facial recognition tool being used by more than 600 US police agencies
22 January 2020
The New York Times published a lengthy profile on Sunday of a company called Clearview AI that has developed a breakthrough facial recognition app reportedly being used by more than 600 law enforcement agencies across the US.
The Times feature, entitled “The Secretive Company That Might End Privacy as We Know It,” is part of a renewed effort within the US political establishment for the adoption of legislative regulation of facial recognition technology that is increasingly and ubiquitously being used by law enforcement.
According to the Times, Clearview AI has been adopted by local police departments, the FBI and the Department of Homeland Security because of its extraordinary ability to determine within minutes the identity of any individual from a single uploaded photo. The Clearview app returns publicly available photos that match the uploaded photo of the individual along with links to the locations on the internet where the public photos appeared.
The Times explains, “The system—whose backbone is a database of more than three billion images that Clearview claims to have scraped from Facebook, YouTube, Venmo and millions of other websites—goes far beyond anything ever constructed by the United States government or Silicon Valley giants.”
According to the report, there are two aspects to this Clearview backbone. The first is a program that crawls the internet and social media platforms, scrapes every face image it can find and records its location. The second is the facial recognition algorithm that converts each face image into a vector face profile—the mathematical relationships within facial geometry such as the distance between the eyes, the distance between the nose and the chin, etc.—and stores these profiles into the searchable database.
On a barebones website, Clearview AI offers marketing language to explain that the system is for law enforcement and is a “technology that helps solve the hardest crimes.” The site also includes a “Request Access” form for prospective police users to sign up for a login. According to the Times, police departments are encouraged to sign up and try the system with a 30-day free trial.
The “Clearview Facts” portion of the marketing information seeks to cover up the invasive and unconstitutional activities of the service by reassuring the public that the system only searches the “open web” for everyone’s photos. It says, “Clearview is an after-the-fact research tool. Clearview is not a surveillance system and is not built like one.”
This dubious explanation ignores the fact that the act of gathering every publicly available facial photo of every person and enabling the state to scan through these photos at will is a gross violation of Fourth Amendment rights against unreasonable searches and seizures.
One of the more frightening revelations of the Times report is that Clearview is developing an interface between the facial recognition app and augmented reality glasses. In theory, such a technology would enable law enforcement to walk down the street or into a crowd or demonstration and every person encountered would be identified in their field of vision in the same way that Facebook tags friends in photos posted to its social media platform.
As is now standard operating procedure from the Pentagon and CIA on down to local sheriff and police departments, the assault on democratic rights is justified by the need to “help to identify child molesters, murderers, suspected terrorists, and other dangerous people quickly, accurately, and reliably to keep our families and communities safe.” Clearview also says that it has been certified as “100% accurate across all demographic groups according to the ACLU’s facial recognition accuracy methodology.”
According to the Times report, Clearview AI was founded in 2016 by Hoan Ton-That, who was born in Australia and moved to Silicon Valley in 2007. Ton-That had been involved in several unsuccessful internet startups before he met Richard Schwartz—a former aide to Rudy Giuliani when he was mayor of New York City between 1994 and 2001—and shortly thereafter founded the company.
Through Schwartz, Clearview AI secured Silicon Valley investment funds of $7 million from billionaire Pay-Pal cofounder Peter Thiel, a prominent supporter of President Donald Trump, among others. Schwartz was also instrumental in leveraging his connections with Republicans to help promote Clearview and present it to police departments across the country.
According to police detectives interviewed by the Times, the Clearview solution is superior to other facial recognition and biometric databases currently available to law enforcement because it contains far more facial images in its database. Plus, users claim, the software can find the identity of individuals from imperfect photos because the algorithm “doesn’t require photos of people looking straight at the camera.”
According to one example provided by the Times, the police department in Clifton, New Jersey “identified shoplifters, an Apple Store thief and a good Samaritan who had punched out a man threatening people with a knife,” during its free trial use period. Frankly, some of the claims made by the company and the users of the Clearview technology sound more like corporate promotional hype than facts.
In any case, it is apparent that several economic and political trends are driving forward the massive development of facial recognition technology and its increased use by the state. A basic Google search of “facial recognition solutions” yields dozens upon dozens of companies that are offering biometric identification systems with various features from surveillance to access control to hardware systems, software systems, database management and analytics.
With the explosion of big data and artificial intelligence technologies for business and government, enormous financial resources are being invested in various facial recognition technologies with the expectation of fantastic returns. According to some estimates, the market for facial recognition systems will reach $7 billion by 2024 and, more generally, AI technologies will reach $126 billion by 2025.
At the same time, the intensification of the class struggle and the growing use of encrypted communications by the public has driven the state into the more aggressive deployment of biometric tools such as facial recognition technologies as a primary means of building up the surveillance apparatus. If the claims about the performance of technologies like Clearview AI are true, the ability of law enforcement to retrieve the identity of anyone within seconds of capturing a photo of them represents a new level of encroachment on the democratic rights of the public.
Like the New York Times, many of the critics of the use of facial recognition tools including the ACLU, accept the use of the technology by the state with the proviso that there be laws governing its use and standards of accuracy established. The timing of the Times article, in fact, coincided with the third facial recognition hearing in the US House of Representatives Oversight Committee on January 15.
The hearing took testimony from nongovernment and government technology experts on the state of the technology and was once again preoccupied with the subject of race and gender bias in face recognition algorithms. None of the speakers called for the use of facial recognition by the police and military to be stopped immediately.
Typical of the outlook of both Democrats and Republicans in Congress, Rep. Gerry Connolly (Democrat from Virginia) said, “We’re going to have to really grapple with what are the parameters of protecting privacy and controlling the use of this technology? Irrespective of its accuracy, there are intrinsic concerns with this technology and its use.”