The promise and hazard of facial biometrics played out in one New York startup
Anyone wondering what the future of biometric facial recognition looks like in the United States would have a good model in a startup reported on late last week by The New York Times. That future looks as bright as some have predicted, as scary as others have worried and very much here.
The company, based in New York City, is Clearview AI. It matches photographs submitted by subscribers against more than 3 billion photographs scraped from uncounted public Web sites including those of social media, according to reporting by The Times.
Clearview AI’s founder, an Australian of Vietnamese descent, is Hoan Ton-That. Ton-That’s personal web site identifies him as co-founder and CEO of Clearview. His co-founder is Richard Schwartz, an aide to Rudolph Giuliani back when Giuliani was New York City mayor.
Ton-That has made multiple stabs at entrepreneurial success, but this is the first business of his to apparently take off. According to the article, at least 600 law-enforcement agencies — federal, state and local, here and abroad, have used Clearview AI’s technology.
Capital-markets publisher Pitchfork reports that the company’s most recent round of financing was a series A traunch totaling $7 million. It has two financial backers, according to Pitchfork.
One is controversial venture capitalist Peter Thiel, famous for backing Facebook and infamous for personally bankrupting Gawker, a free-wheeling blog network that ran afoul of Thiel. Venture-capitalist firm Kirenaga Partners is the second to put money behind Ton-That’s company. Kirenaga’s holdings cluster around health-care and edge firms.
Success stories by police officers using Clearview AI are compelling and have played a significant — but not sole — role in generating new business. The Times and other news publications cite the example of a case solved by the Indiana State Police within 20 minutes of starting the application. Other similar successes are touted by the company.
Clearview AI’s client roster also has benefited from Schwartz’s participation. His connections, according to The Times, are voluminous thanks to his time with Giuliani and his stint as editorial page editor of The New York Daily News.
The Times story is raising red flags. As a privately funded company, there is no way to know what kind of protections — if any — there are for people who find themselves among billions of others in a photo database open to police, yes, but any other agency, company or individual clients that Ton-That chooses to sign up.
At the moment, according to the article, there is no way to even ask Ton-That to remove one’s photograph from his database. That includes shots that police are uploading to be matched.
Underscoring the concern is the by-now well-known fact that facial recognition technology is not foolproof, especially when it comes to images of people of color. While Ton-That shared success stories with The Times, there was no mention of false positives that have occurred, a prospect that grows with the size of a database.
biometric database | biometrics | Clearview AI | dataset | facial recognition | police | privacy