What a drag(net)
Sacramento sheriff’s deputies use facial recognition technology to solve crime of walking with an open alcohol container
Facial recognition technology is in the hands of Sacramento law enforcement—and it isn’t necessarily being used to catch the next Golden State Killer.
On the morning of Nov. 8, 2019, Sacramento County sheriff’s deputies arrested a 26-year-old African-American woman in South Sacramento for carrying an open alcoholic beverage and for allegedly lying about her identity when they stopped her.
According to an arrest summary, one of the deputies stopped the woman because she was walking northbound on Stockton Boulevard in North Highlands, “a known prostitution stroll.”
The woman allegedly identified herself to the deputy by another name, but didn’t resemble that woman’s DMV photo, the arrest summary states. “Officers used facial recognition to determine” her real identity, then cited and released her 80 minutes later for the open container infraction and false ID, the summary states.
The Sheriff’s Office contracts with a company called DataWorks Plus to capture and store mugshots, crime scene photos and other digital images. Since late 2017, the Sheriff’s Office has been able to run its growing trove of images through a facial recognition database that DataWorks created for California law enforcement agencies, including some of the state’s largest.
According to an October 2017 staff report to the Sacramento County Board of Supervisors, DataWorks created the California Facial Recognition Interconnect system, or CAFRI, to allow participating law enforcement agencies to match photos of an individual against existing photos in the CAFRI system.
At that time, the Sacramento County Sheriff’s Office was the sixth agency to join CAFRI, which already counted Los Angeles, Riverside, San Bernardino, San Joaquin and San Diego counties as subscribers.
According to the letter of agreement that the Sheriff’s Office signed, DataWorks only makes CAFRI available for law enforcement agencies conducting law enforcement business. The company does allow for the possibility of adding out-of-state mugshot customers, but says that no private agencies can access its data or images. “DataWorks Plus will not sell information or data to private parties that reside on this server,” the agreement letter states.
That lends the company some distinction from Clearview AI, a secretive start-up company whose founder hasn’t sworn off marketing his internet-scraping app to private corporations and foreign governments.
“There’s a fundamental difference between the two,” said Dave Maass, a senior investigative researcher with the Electronic Frontier Foundation, a digital rights nonprofit based in San Francisco. Where DataWorks stockpiles images of people who have been arrested, Clearview’s dragnet encompasses “basically anybody who’s ever had an image of themselves online,” Maass added.
Hundreds of law enforcement agencies already use the Clearview AI app. The New York Times uncovered the chilling privacy threat in January, prompting some cities and states to swear off the technology. Facebook and YouTube have also ordered Clearview AI to desist from scouring their users’ data, though it’s unclear to what effect.
As for whether facial recognition technology is being used to solve more serious crimes in Sacramento County, a sheriff’s spokeswoman couldn’t be reached before deadline. But Maass said the open beverage arrest wouldn’t be atypical. While law enforcement agencies tell elected officials they need this technology to combat domestic terrorism, human trafficking and violent crimes, “it ends up being used in these very mundane ways,” Maass said.
Gov. Gavin Newsom signed a bill into law prohibiting law enforcement from using facial recognition in body cams and placing a three-year moratorium on handheld devices.
The Sacramento County Sheriff’s Office has yet to provide deputies with body-worn cameras, despite recommendations by civil liberties groups and the county’s former inspector general, whom Sheriff Scott Jones forced out in 2018.
This story has been updated with the correct neighborhood location.