Around 5 p.m. on May 20, 2019, a teller at the Call Federal Credit Union in Chesterfield, Virginia, got ready to help her next customer. He gave her a handwritten note: “I’ve been watching you for sometime now. I got your family as hostage and I know where you live,” it said, according to a court brief. “If you or your coworker alert the cops or anyone your family and you are going to be hurt… I need at least 1OOk.”
The teller told him she didn’t have access to that kind of money, at which point he pulled out a silver and black handgun. Waving it around, he herded the bank’s customers and employees behind the counter and into the back room. There he ordered everyone to their knees and forced the manager to open the safe. He fled with $195,000.
No one recognized the thief, although a witness at a church near the burglary reported having seen a man who looked suspicious in a blue Buick sedan, the court brief said. The police had little else to go on, except surveillance tape showing him entering the bank with a cell phone to his ear and passing by that church. Three and a half weeks later, still empty-handed, they asked a magistrate for a geofence warrant.
Unlike a traditional search order that identifies a particular suspect, geofence warrants require Google to trawl its massive library of location data, commonly known as the “Sensorvault,” to identify people who were in the area when a crime was committed. They are relatively new, and increasingly widespread: Between 2017 and 2018, Google saw a 1,500% surge in the number of requests it received, and from 2018 to 2019, the rate increased over 500%.
The resulting data helped police solve a murder case in Cobb County, Georgia, and to identify suspects in a home invasion in Eden Prairie, Minnesota, a fatal shooting in Raleigh…
