Clearview AI Does Well in Another Round of Facial Recognition Accuracy Tests

Ad Blocker Detected

Our website is made possible by displaying online advertisements to our visitors. Please consider supporting us by disabling your ad blocker.

After Clearview AI scraped billions of photos from the public web — from websites including Instagram, Venmo and LinkedIn — to create a facial recognition tool for law enforcement authorities, many concerns were raised about the company and its norm-breaking tool. Beyond the privacy implications and legality of what Clearview AI had done, there were questions about whether the tool worked as advertised: Could the company actually find one particular person’s face out of a database of billions?

Clearview AI’s app was in the hands of law enforcement agencies for years before its accuracy was tested by an impartial third party. Now, after two rounds of federal testing in the last month, the accuracy of the tool is no longer a prime concern.

In results announced on Monday, Clearview, which is based in New York, placed among the top 10 out of nearly 100 facial recognition vendors in a federal test intended to reveal which tools are best at finding the right face while looking through photos of millions of people. Clearview performed less well in another version of the test, which simulates using facial recognition for providing access to buildings, such as verifying that someone is an employee.

“We’re pleased,” said Clearview’s chief executive, Hoan Ton-That. “It reflects our actual-use case.”

The company also performed well last month in a test — called a one-to-one test — of its ability to match two different photos of the same person, simulating the facial verification that people use to unlock their smartphones.

The positive results have “been a shot in the arm for the sales team,” Mr. Ton-That said.

The National Institute of Standards and Technology has been administering Face Recognition Vendor Tests for two decades. Since those tests began, the report notes, “face recognition has undergone an industrial revolution, with algorithms increasingly tolerant of poorly illuminated and other low-quality images, and poorly posed subjects.”

Clearview made an impressive debut on the charts for investigative, or one-to-many, searches, but the top performers were SenseTime, a Chinese company, and Cubox, from South Korea. In 2019, the Commerce Department blacklisted SenseTime and 27 other Chinese entities because their products were implicated in China’s campaign against Uyghurs and other Muslim minorities. Axios has reported that the designation was later changed to “Beijing SenseTime,” limiting the effects of the blacklisting.

Accuracy aside, questions remain about the legality of Clearview’s tool. The authorities in Canada and in Australia have said Clearview broke their laws by failing to get the consent of citizens whose photos are included in the database, and the company is fighting lawsuits over privacy in Illinois and Vermont.