CLEAR AND PRESENT DANGER

Clearview Al's far-right founders always intended to target immigrants and the political left. Now their facial recognition dragnet is in the hands of the Trump administration.

- Luke O'Brien

Hoan Ton-That, an Australian coder building a powerful facial recognition system, emailed his American business partners with a plan to deploy their fledgling technology. "Border patrol pitch," the subject line read. He hoped to persuade the federal government to integrate their product with border surveillance cameras so that their newly formed company, later named Clearview Al, could use "face detection" on immigrants entering the United States.

An immigrant to the United States himself, Ton-That grew up in Melbourne and Canberra and claimed to be descended from Vietnamese royalty. At 19, he dropped out of college and, in 2007, moved to San Francisco to pursue a tech career. He later fell in with Silicon Valley neoreactionaries who embraced a far-right, technocratic vision of society. Now Ton-That and his partners wanted to use facial recognition to keep people out of the country. Certain people. Their technology would put that ideology into action.

Clearview had compiled a massive biometric database that would eventually contain billions of images the company scraped off the internet and social media without the knowledge of the platforms or their users. Its AI analyzed these images, creating a "faceprint" for every individual. The company let users run a "probe photo" against its database, and if it generated a hit, it displayed the matching images and links to the websites where they originated. This made it easy for Clearview users to further profile their targets with other information found on those webpages: religious or political affiliation, family and friends, romantic partners, sexuality. All without a search warrant or probable cause.