Home News She told the BBC that US police had used Clearview AI nearly a million times

She told the BBC that US police had used Clearview AI nearly a million times

by admin
  • Written by James Clayton and Ben Derico
  • BBC News, San Francisco

image sources, Spencer Wallen / Eye M

The founder of facial recognition company Clearview says that nearly 1 million searches have been made by the US police

CEO Hwan Tun also revealed that Clearview now has 30 billion photos deleted from platforms like Facebook, taken without users’ permission.

The company has been repeatedly fined millions of dollars in Europe and Australia for breaching privacy.

Critics argue that police use of Clearview puts everyone in a “perpetual police line”.

“Every time they have a picture of a suspect, they compare it to your face,” says Matthew Guaragilia of the Electronic Frontier Foundation. “She’s very aggressive.

The Clearview system allows a law enforcement officer to upload a photo of a face and find matches in a database of the billions of photos they have collected.

It then provides links to where the corresponding images appear on the Internet. It is considered one of the most powerful and accurate facial recognition companies in the world.

image sources, Spencer Wallen / Eye M

Photo caption

Hwan Tun That, founder and CEO of Clearview AI, speaking to the BBC

The company is barred from selling its services to most US companies, after the American Civil Liberties Union (ACLU) sued Clearview AI in Illinois for violating the Private Life Protection Act.

But there is an exception for the police, Ton says — its software is used by hundreds of police forces across the United States.

Police use of facial recognition is often sold to the public as being used only for serious or violent crimes.

In a rare interview with law enforcement about Clearview’s effectiveness, Miami police said they have used the software for all kinds of crimes, from murders to shoplifting.

Deputy Police Chief Armando Aguilar said his team uses the system about 450 times a year and it has helped solve many murders.

However, critics say there are almost no laws on police use of facial recognition.

Photo caption

Miami Police Deputy Chief Armando Aguilar

Miami police treat facial recognition as tip-off, Aguilar says. “We don’t give up because the algorithm tells us to,” he says. “Either we put this name in a collage, or we start solving the issue through traditional means.”

false identity

There are a few documented cases of misidentification by facial recognition by the police. However, the lack of data and transparency about police use means the real figure is likely to be much higher.

Ton says he is not aware of any cases of misidentification using Clearview. He agrees that the police made fake arrests using facial recognition technology, but blames it on “police weakness”.

Clearview often cites research that shows an accuracy rate close to 100%. But these numbers are often based on mugshots.

In fact, Clearview’s resolution depends on the quality of the image being fed in — something Mr. Ton-That agrees with.

Civil rights activists want police forces to use Clearview to say openly when they use it — and publicly test its accuracy in court. They want the algorithm to be reviewed by independent experts and are skeptical of the company’s claims.

Caitlin Jackson is a New York-based criminal defense attorney who campaigns against the use of facial recognition technology by the police.

“I think the truth is, the idea that this is incredibly accurate is wishful thinking,” she says. “There is no way to tell if you are using wild images like CCTV screenshots.”

Photo caption

Caitlin Jackson, a defense attorney in New York

However, Ton-That told the BBC he did not want to testify in court about its accuracy.

“We really don’t want to be in court to testify about the accuracy of the algorithm…because investigators also use other methods to check it,” he says.

Ton says he recently gave the Clearview system to defense attorneys on select cases. He believes that prosecutors and lawyers should have equal access to technology.

Last year, charges against Andrew Conlin of Fort Myers, Florida, were dropped after he used Clearview to find a key witness.

Conlin was a passenger in a friend’s car in March 2017 when it crashed into palm trees at high speed.

The driver was thrown from the car and killed. Mr. Conlin was pulled from the wreckage by a bystander, but left without making a statement.

Although Mr. Conlin said he was the passenger, the police suspected him of driving and charged him with murder.

His lawyer had a photo of a bystander taken from police camera footage. Just before his trial, Mr. Ton-That allows Clearview to be used in the case.

Christopher O’Brien, Conlin’s defense lawyer, told the BBC: “The AI ​​appeared in three to five seconds. “It was phenomenal. »

Photo caption

Andre Conlin

The witness, Vince Ramirez, testified that he removed Mr. Conlin from the passenger seat. Shortly thereafter, the charges were dropped.

But although there are cases when Clearview has proven itself, some believe that its price is too high.

“Clearview is a private company that bioprints people’s faces from their photos online without their consent,” says Guaragilia.

“It’s a huge civil liberties and civil rights issue, and it desperately needs to be banned.” »

UK viewers can watch the Our World documentary at Clearview AI on BBC iPlayer

Related News

Leave a Comment