Clearview AI is forbidden from selling to most US companies as a result of a ruling by an Illinois court, when the American Civil Liberties Union (ACLU) took Clearview AI for violating privacy law, but US police are an exception. 

Clearview allows police to upload a photo of someone's face and find matches of it in a database of billions of images, then it provides a link to an online presence of that match - Clearview is said to be one of the globe's most powerful yet accurate facial recognition services.

The CEO of the company, Hoan Ton-That, argues that hundreds of law enforcement in the US use the service, even though it is prohibited in the cities of Portland, San Francisco, and Seattle.

The Assistant Chief of Miami's Police force, Armando Aguilar, confirmed that his unit used Clearview to identify suspects of various crimes, adding that it was used nearly 450 times a year. 

Read more: UN says AI poses 'serious risk' for human rights

"We don't make an arrest because an algorithm tells us to," Aguilar said, adding: "We either put that name in a photographic line-up or we go about solving the case through traditional means."

However, it's not always 100% accurate as there have been cases of mistaken identity, but Ton-That claims he is unaware of any mistaken identity incidents with his company and instead blames 'poor policing' for wrongful arrests and mistaken identities after using facial recognition. 

'Accuracy is wishful thinking'

Civil rights campaigners want police forces that use Clearview to openly say when it is used - and for its accuracy to be openly tested in court. They want the algorithm to be scrutinized by independent experts and are skeptical of the company's claims.

Kaitlin Jackson, a New York criminal defense lawyer against police use of facial recognition states, "I think the truth is that the idea that this is incredibly accurate is wishful thinking," she says. "There is no way to know that when you're using images in the wild like screengrabs from CCTV."

But Ton-That argued that he doesn't want to testify about "the accuracy of the algorithm… because the investigators, they're using other methods to also verify it." 

Read next: ChatGPT bug results in breach of personal user data

One case last year is an exhibit of this case. Andrew Conlyn from Florida was a passenger in a friend's car that crashed at high speed in March 2017. The driver was killed and someone was able to pull Conlyn from the car but left immediately. 

Despite the fact that Conlyn told police he was the passenger, they suspected he was the driver and charged him with vehicular homicide. When his lawyers were able to show an image of the passerby that pulled him out of the car from police cam footage, Ton-That gave the green light for Clearview to be used. 

"This AI popped him up in like, three to five seconds," Conlyn's lawyer, Christopher O'Brien, told the BBC. "It was phenomenal."

Not only did the passerby make a statement, but police dropped charges against Conlyn. 

"Clearview is a private company that is making face prints of people based on their photos online without their consent," said Matthew Guaragilia from the Electronic Frontier Foundation. "It's a huge problem for civil liberties and civil rights, and it absolutely needs to be banned."