The New York Times exposes Clearview AI

Reporter Kashmir Hill’s New York Times investigation revealed that a small, previously unknown startup, Clearview AI, had built a facial recognition tool by scraping roughly three billion photos from Facebook, YouTube, Venmo and other public sites without the platforms’ or subjects’ consent. Uploading a photo to the app returned matching images of the same person and links to where they had been posted online — a capability Google and other large companies had previously judged too dangerous to release. The article reported that Clearview had quietly sold access to more than 600 law enforcement agencies, including the FBI and the Department of Homeland Security, and to some private companies, with little public disclosure of who was using it or how.

The report prompted an immediate response from the platforms Clearview had scraped. Within weeks, Google, YouTube, Venmo, LinkedIn and Twitter sent cease-and-desist letters demanding it stop collecting their users’ images and, in Twitter’s case, delete data already taken; Facebook sent a similar demand without issuing a formal cease-and-desist. Clearview’s founder, Hoan Ton-That, argued the company had a First Amendment right to index publicly available data, a defence that became central to the lawsuits that followed, including a biometric-privacy case in Illinois that Clearview later settled.

The story is credited with moving facial recognition from a specialist policy concern into mainstream controversy, establishing that the technology’s practical limit was not what companies were willing to build but what they were willing to admit to using — and setting up a year in which several major vendors abandoned or paused police sales of the technology altogether.