California Skeptical of Police Using Facial Recognition Software

By: David Cruz Quevedo

In May, after San Francisco became the first city and county to ban the use of facial recognition software by law enforcement, Oakland, Berkeley and Somerville, Massachusetts followed suit. In October, California Governor Gavin Newsom approved a ban on police use of facial recognition software for body cameras. 

The American Civil Liberties Union of Northern California (ACLU-NC) led the campaign that resulted in the San Francisco ban, having stated that this technology reinforces biases against people of color. Matt Cagle, a technology and civil liberties attorney with ACLU-NC, said law enforcement’s use of this technology is concerning. 

“Many prominent facial recognition systems suffer from accuracy issues and bias issues because they are able to recognize White faces generally better than they can recognize Black and Brown faces,” Cagle said. 

For example, he explained, earlier this year the ACLU conducted a test with Amazon’s facial surveillance technology, called “Rekognition,” and found that it incorrectly matched 28 members of Congress with mugshots of other people who had been arrested for a crime. These false matches were disproportionately people of color, including the civil rights icon, Rep. John Lewis, D-G.A., and five other members of the Congressional Black Caucus. 

Cagle compared the use of facial recognition technology with the Supreme Court’s 2018 ruling in Carpenter v. U.S. Cagle summarized the Court’s opinion in Carpenter saying, “the government needs to get a warrant to obtain long-term historical locations from the cell phone carrier.” The Court rejected the idea that “simply because you go into society, you somehow give up your rights to privacy,” Cagle said. “We think facial recognition technology is significantly more dangerous than cell phone location searches because if the government implements facial recognition network, they don’t need to go to the cell phone carrier.” 

The use of facial recognition technology is dependent on a number of factors. First, a camera must capture a useable picture. Then, the image is run through software that detects unique physical features like distance between the eyes and nose length. Once those key facial features have been determined, the software compares the image against a large database of photos in hopes of getting a match. From there, the algorithms generate potential matches. Then, a human operator scans through the suggestions to make a final determination.

Supporters of facial recognition technology look to the potential benefits that this new innovation can have on public safety. James Leonard, Deputy District Attorney for Santa Clara County, suggested that police departments will use “facial recognition [as] a lead [to point] an investigator in the direction to go.'' 

Leonard offered an example of a New York police department that was able to arrest a man in 24 hours who had nearly raped a woman at knife-point due to the use of facial recognition technology. 

With facial recognition technology, “we have a situation where someone can be misidentified, but if we are going to throw it out just because someone can be misidentified then we would have to throw out eyewitness identification too because people misidentify people too,” Leonard said. 

Leonard also said he has not found a single case that has been convicted using facial recognition identification alone.  

According to a research report by MarketsandMarkets, a global market research and consulting firm, law enforcement use of facial recognition is expected to help the industry grow from $3.2 billion in 2019 to $7 billion in 2024. 

As the industry is expected to grow, most Americans don’t seem too concerned with the implications. A Pew Research study from September found 56 percent of Americans trust law enforcement to use facial recognition technology responsibly. However, in a March study conducted in California by the ACLU-NC, 82 percent of respondents said they would be opposed to the government using facial recognition technology. 

(Editor's Note: This article was originally published in the December 2019 [Volume 50, Issue 2] version of The Advocate.)