San Francisco is expected to set a groundbreaking precedent on Tuesday by voting to become the first city in the country to ban police from using facial recognition.
Part of the reason: concerns about accuracy.
“With Caucasian faces, facial recognition is pretty good. It has a 90 to 95 percent accuracy rate,” explains Darrell West, director of the Center for Technology Innovation with the Brookings Institution. “But with minorities, sometimes the accuracy rate drops to 70 percent.”
West also says that once a person’s image is in the database, there’s uncertainty surrounding what it could be used for.
A Georgetown law study found 1 in 2 American adults is in a law enforcement face recognition network.
Law enforcement has argued the technology helps solve crimes or improve investigations. Agencies across the country can use driver’s license photos or mug shots to match someone's identity.
“All it's doing is using something that's readily available,” says Sheriff Bob Gualtieri with the Pinellas County Sheriff’s Department in Florida.
But it's not just law enforcement using the technology. Stores, airports and some concert venues are all starting to work it into their operations.
It's become so mainstream, Congress is now considering a bill to stop businesses from collecting facial recognition data on customers without their permission.
“I think people find it very intrusive that you're just walking down the street or going into the store and somebody's recording your face and then attaching your identity to that image,” West says.
If the bill in Congress passes, it would be the first federal law on facial recognition.