Flock Safety Develops AI to Track Drivers by Movement Alone
Flock Safety, a surveillance company whose cameras are deployed in over 6,000 U.S. communities to log vehicle movements, has developed an artificial intelligence tool capable of identifying and tracking drivers based solely on their travel patterns.
The existence of this system was revealed in a report by Wired, which obtained the software’s code from files accessible on Flock’s own website.
The software, initially named Nightshift and now referred to as OS Investigate, is currently in development and undergoing testing with a select group of law enforcement agencies, according to the company.
The system allows officers to select from a menu of 69 pre-written prompts or to input custom queries. One such prompt is designed to identify “witnesses,” listing vehicles most frequently observed in a neighborhood over the preceding two weeks. Another query can identify individuals arrested more than twice in a two-year period (excluding drug offenses) and compile profiles on the top three suspects. A “workup” function can generate a dossier on an individual, including relatives, phone numbers, and online accounts, by cross-referencing police records with commercial data brokers, using just a name and date of birth.
Crucially, 14 of these prompts do not require any vehicle plate number, name, or description. Instead, an officer can specify a location, a timeframe, and a behavior—such as cars involved in incidents at three retail stores over three days, or multiple bank visits within a week—and the system will return individuals matching the criteria. Filters are in place to exclude buses and delivery vans, ensuring that ordinary drivers performing routine tasks remain on the list.
The underlying logic for identifying associates is detailed in the code. It works by counting how often other license plates appear at the same cameras within a two-minute window of a target vehicle. Any plates appearing three or more times with a confidence score above 0.75 are flagged and up to 20 associated names are returned.
“I don’t know how else to say this, but this sounds completely insane,” Noel Pichardo, a former police officer in Pawtucket, Rhode Island, who reviewed the prompts, told Wired. “I don’t know how anyone can argue against the idea that Flock literally tracks people.”
Jay Stanley, a senior policy analyst at the ACLU, expressed concern that the capabilities of Flock’s cameras, combined with this new software, create a surveillance environment with “very little space… like China.”
Flock has consistently stated that its technology “cannot recognize, identify, or track individuals” and is “not for watching people.” Spokesperson Paris Lewbel described OS Investigate as a distinct product from the company’s license-plate readers, designed to assist investigators in working with existing agency records. Lewbel also noted that its capabilities may undergo significant changes before any broader release. The company did not dispute Wired’s reporting.
The controversy has already led to direct action. In June and July, activists vandalized, covered, or destroyed Flock cameras across the country, with The Guardian documenting at least 33 incidents in 23 states. This backlash is consistent with documented instances of police misuse of the network. A report by the Institute for Justice found over two dozen cases where officers allegedly used license-plate readers to stalk current or former partners, leading to resignations or arrests.
The issue has also reached Capitol Hill. Representative Thomas Massie has proposed legislation that would prohibit federal funding for agencies deploying this technology.
Meanwhile, researchers have developed AI-generated patterns designed to confuse the plate-reading software. Flock is not the only entity pushing the boundaries of surveillance technology; Meta has patented cameras capable of recognizing faces and logging individuals’ actions.
