Clearview’s AI may be able to recognize millions of faces thanks to images uploaded to social media

Clearview AI is a controversial company. Even though it was banned from selling its services to companies, the facial recognition software is still used in thousands of instances related to police investigations and law enforcement.

Company CEO Hoan Ton-That sat down for an interview with the BBC, where he disclosed that the company’s database contains over 30 billion images. How did they get them? Well, through the public profiles of users on social media websites like Facebook, of course.

That isn’t to say that Clearview has gotten away with this. The company has been fined numerous times for millions of dollars by countries in Europe. That, however, hasn’t stopped US police from utilizing the facial recognition software, which have now reached the million mark.



Now, law enforcement hasn’t confirmed the claims made by Clearview, but still — a million searches is a concerning number. The Miami Police Department, for example, has been transparent about using the software around 450 in the last year. That being said, they’ve also gone on record saying that they don’t arrest people just because the computer tells them that it has found a match. In other words: Clearview is just another tool in the box.

But precisely because very few law enforcement agencies have come forward with factual usage statistics, the success rate of Clearview is near impossible to track. The company’s CEO claims that he isn’t aware of an instance where the software failed, but is that really a trustworthy claim? What about the instances where the software was a key factor in cracking a difficult and high-profile case —  do we know how useful it actually is?