The troopers’ use of the technology was first reported by BuzzFeed News on Thursday afternoon. BuzzFeed obtained a list of all of Clearview AI’s clients.
Facial recognition isn’t new technology, however the way Clearview AI has reportedly used it is.
Clearview AI gained attention after a New York Times investigation revealed that the company scraped billions of images from social media networks like Facebook, Instagram and YouTube. It often gives free trials to law enforcement agencies, who then buy licenses to use the app, the Times reported. Those agencies are then able to match images of unknown people with pictures online.
On Wednesday, The Daily Beast obtained a memo sent to Clearview AI’s customers acknowledging its entire client list was breached. The memo, however, said its clients’ search history had not been accessed.
Previously, Clearview AI told the New York Times that its client list included 600 local law enforcement agencies. BuzzFeed News’ report Thursday said Clearview AI’s client list actually has 2,900 clients, including companies like Macy’s and Best Buy as well as federal law enforcement. Immigration and Customs Enforcement, the FBI, Customs and Border Patrol and the U.S. Attorney’s Office for the Southern District of New York were also named as clients in the BuzzFeed report.
About 2,200 of those clients have searched Clearview AI within the last year, according to BuzzFeed.
Clearview AI’s scraping of images is often against the terms of service set forward by the social media networks it takes the pictures from. Google, YouTube, Venmo, LinkedIn and Facebook have all called for Clearview AI to stop taking images from their sites, according to a CBS News report.
The New York State Police spokesperson said that use of the app is restricted to “trained, authorized members” and is “governed by a strict written policy and closely supervised.” State police paid $15,000 for its licenses, according to BuzzFeed.
Clearview AI told the Times that its app finds matches up to 75 percent of the time, though the app has yet to be tested by any independent organizations, like the National Institute of Standards and Technology.
In December, NIST released a report that found current facial recognition technology came up with false positives more often for Asian and black people when compared with false positives for white people.
NIST did note that different algorithms produce different results and not all the facial recognition algorithms they reviewed had discrepancies.
©2020 Syracuse Media Group, N.Y. Distributed by Tribune Content Agency, LLC.