Home Trending Study finds police have used facial recognition technology on 117 million Americans

Study finds police have used facial recognition technology on 117 million Americans

111
0
SHARE

A body camera is worn by a Miami-Dade Police Department officer.
Image: AP Photo/Lynne Sladky

Police departments across the country may know what you look like, even if you don’t know how they got your photo.

About half of American adults are already in a “law enforcement face recognition network,” according to a new study published Tuesday by researchers at Georgetown University’s Center on Privacy and Technology.

“This technology is not limited to serious criminals,” Alvaro Bedoya, one of the study’s co-authors, said during a conference call on Tuesday. “It’s not limited at all, really.”

A Denver Police Department officer uses a camera to shoot video of bystanders.

Image: AP Photo/David Zalubowski

To many, perhaps the most surprising thing about the report is that it found “face recognition is neither new nor rare.”

A few other findings in the study:

  • Around 117 million American adults are already in a facial recognition network.

  • The FBI runs searches of face recognition databases more often than wiretaps.

  • About 25 percent of police departments across the country have access to facial recognition networks. Those networks are often cross-referenced with databases of ID photos such as driver’s licenses.

And for now, law enforcement can do almost whatever they want with this technology, including scanning the photos of people who have never committed a crime. No state legislature “has passed a law comprehensively regulating police face recognition,” according to the report.

Bay City Police Capt. Joseph Lavana plugs a Taser body-worn camera to the docking station at the police station.

Image: Yfat Yossifor/The Bay City Times via AP

“Face recognition is fundamentally changing the way not just how police interact with the public, but whether and how individuals act in public space,” Neema Singh Guliani, who works in legislative counsel with the ACLU with a focus on surveillance and privacy, said during the call.

The civil liberties issues are profound and may well intensify in years to come as video technology develops.

“Major police departments are exploring real-time face recognition on live surveillance camera video,” according to the report. This technology is often highly inaccurate, because cameras do not often capture images of people head-on, and light surrounding cameras is in constant flux. Video is often poor in quality, and uploading dozens of livestreams at once requires incredible processing power that’s not likely to be readily available to many departments.

But at least five “major police departments,” including departments in Los Angeles, Chicago and Dallas, have either already used the technology, bought it from a vendor, or have said they’re interested in acquiring it.

Police use of facial recognition also has the potential to “supercharge” racist policing, according to Sakira Cook, who took part in the call and works in public policy at The Leadership Conference on Civil and Human Rights. The report backs up her words.

“Because African-Americans are more targeted by law enforcement, it’s more likely this technology will be used on them,” Bedoya said.

Black Lives Matter protesters march in downtown Atlanta, on Sept. 24.

Image: AP Photo/Branden Camp

A separate study co-authored by the FBI also showed facial recognition technology is less accurate when trying to identify black people, meaning innocent black people are disproportionately likely to be suspected for a crime they did not commit, especially considering black people are already disproportionately arrested by police.

Companies that develop facial recognition admitted they didn’t test for racial bias, and law enforcement groups have done little to correct the error.

“We know of only two departments who actually set an accuracy standard for the systems they purchased,” Jonathan Frankle, another of the study’s co-authors, said during the call.

Law enforcement departments try to plug the technology’s accuracy gaps with actual officers, but Frankle said studies have shown that a person’s ability to identify individuals fades when trying to identify those of a different race.

The nearly invisible proliferation of facial recognition technology is on par with the nearly invisible proliferation of other surveillance technologies used by law enforcement throughout the country.

Police departments from Baltimore to Fresno, California, have used social media surveillance software that allows officers to follow tweets, Instagrams and Facebook posts of individual users.

Geofeedia, a company used by media and law enforcement, marketed its social media surveillance software to law enforcement as a means to track activists and protesters, but its access to data from Facebook, Twitter and Instagram was cut after an ACLU investigation.

Police guard an intersection on May 4, 2015, in Baltimore.

Image: Algerina Perna/The Baltimore Sun via AP

Departments have also used devices called “Stingrays” to capture cellphone data from nearby people without their knowledge.

The authors of the Georgetown report hope politicians soon craft laws to curb how law enforcement uses facial recognition.

They recommend not allowing facial recognition software to mine databases of driver’s licenses that contain photos of people who have never committed a crime, and they hope real-time facial recognition will only be used after officers “specify where continuous scanning will occur, and cap the length of time it may be used.”

They also recommend companies designing facial recognition software test their systems for racial bias and make the results of those tests available to the public.

For now, the machinations behind how and when police use facial recognition to monitor the public remain largely unknown.

“I think we have to ask ourselves, does that look like America?” Bedoya said. “We don’t think it does.”

Read more: http://mashable.com/2016/10/18/police-facial-recognition-surveillance/