Flock Safety AI Tracks Drivers by Behavior, Not Just Plates

Surveillance Expands Beyond License Plates

Flock Safety, whose automated license plate readers operate in more than 6,000 U.S. communities, has developed artificial intelligence software capable of tracking individuals based purely on their driving patterns—no plate number, name, or vehicle description required.

The system, currently named OS Investigate and previously called Nightshift, remains in development with a limited group of law enforcement testing partners. Details of its capabilities emerged after Wired extracted the tool's code from files publicly accessible on Flock's login pages.

How the Pattern-Matching Works

Officers interact with OS Investigate through a chat interface offering 69 pre-written queries. Fourteen of these prompts require no identifying information at all. Instead, an officer enters a location, timeframe, and behavioral pattern—such as vehicles visiting multiple retail stores over three days or hitting several banks within a week. The system then returns matching drivers, filtering out commercial vehicles like buses and delivery vans to isolate private cars.

The software's associate-finding function uses camera timestamps to build networks. When other vehicles appear at the same camera locations within two minutes of a target car, the system flags those that cross paths three or more times with at least 0.75 confidence, returning up to 20 associated names.

Other built-in queries generate "witness" lists of vehicles most frequently seen in a neighborhood over two weeks, or compile dossiers on individuals arrested more than twice in two years—excluding only drug offenses. A "workup" feature converts a name and birthdate into relatives, phone numbers, and social media accounts by pulling from police databases and commercial data brokers.

Why It Matters

This capability represents a fundamental shift from passive vehicle tracking to active behavioral surveillance. The technology enables law enforcement to identify people not because they're suspected of a specific crime, but because their routine movements match a pattern an officer defines. That inversion—searching for people who fit a behavior rather than investigating known suspects—raises questions about mass surveillance that existing legal frameworks weren't designed to address. For enterprises building location-based services or mobility platforms, Flock's approach demonstrates how aggregated movement data can reveal individual identities even when systems claim not to track people.

Pushback and Policy Response

Noel Pichardo, a former Pawtucket, Rhode Island police officer who reviewed the prompts, told Wired the system amounts to literal people-tracking. Jay Stanley, senior policy analyst at the ACLU, said the gap between Flock's stated limitations and actual capabilities leaves "very little space... like China."

Activists have responded with direct action. Between June and July, The Guardian documented at least 33 incidents across 23 states where Flock cameras were spray-painted, covered, or destroyed. The Institute for Justice found over two dozen cases of officers resigning or facing arrest for allegedly using license plate readers to stalk current or former romantic partners.

Rep. Thomas Massie is preparing legislation to withhold federal funding from agencies deploying the technology.

Flock spokesperson Paris Lewbel characterized OS Investigate as separate from the company's license plate readers, designed to help investigators work with records their agencies already possess. She noted capabilities may change significantly before broader release and did not dispute Wired's technical findings. Flock has consistently stated its technology "cannot recognize, identify, or track individuals" and is "not for watching people."

The details were first reported by Wired.