Flock Safety Cameras: Vehicle Tracking, Data Privacy & Legal Pushback

Flock Safety, an $8.4 billion surveillance startup, has developed a system that uses small black cameras to capture license plate information, process it through an on-device neural network, and upload the results to a nationwide searchable database. This database allows law enforcement across the country to track vehicle movements without a warrant.

How Flock Cameras Work

Flock Safety was founded in 2017 by Garrett Langley, an engineer who sought to address the lack of police investigation into neighborhood crimes without license plate numbers. He initially built an automated license plate reader and sold it to homeowners associations before expanding to police departments. Today, Flock cameras operate in thousands of communities, scanning billions of plates monthly. The company claims its technology helps solve 20% of reported crimes, a statistic that has been met with skepticism regarding its verification methods.

The flagship camera, known as the Falcon, is described as a sophisticated device. It functions as a solar-powered spy device equipped with an LTE modem, infrared night vision, and a motion-triggered camera that can be easily deployed. Instead of streaming video, the Falcon performs machine learning inference at the edge. When a car triggers its motion sensor, an on-device model captures still images and creates a "vehicle fingerprint." This fingerprint classifies various attributes of the vehicle, including:

• Make
• Model
• Color
• Dents
• Rims
• Roof racks
• Bumper stickers

This detailed classification allows the system to identify vehicles even if their license plates are missing or obscured. Only the images and structured metadata are transmitted via LTE to Flock Cloud, which minimizes bandwidth usage and keeps the unit cost-effective for widespread deployment.

Once in the Flock Cloud, each scanned vehicle becomes a record in a database. This database is continuously cross-referenced with "hot lists" for items such as stolen vehicles and Amber Alerts. If a wanted vehicle passes a camera, law enforcement in the vicinity receive push notifications.

The Value of Aggregated Data

While individual scans provide a timestamped license plate, the aggregation of billions of scans creates a comprehensive record of vehicle movements. This data can reveal patterns such as where individuals live, work, and visit. Police departments can opt into nationwide sharing, enabling officers in small towns to query vehicle histories across all 50 states.

Legal Basis: The Third-Party Doctrine

The legality of Flock Safety's operations hinges on a 50-year-old legal principle known as the "third-party doctrine." Established by the Supreme Court in the 1970s, this doctrine states that individuals lose a reasonable expectation of privacy once they voluntarily share their data with a third party. This creates a loophole where the government can bypass Fourth Amendment warrant requirements by accessing data from third-party services.

The system relies on law enforcement officers to input a reason for their searches. However, this honor system has proven to be vulnerable to abuse. Examples include:

• An Idaho sheriff running his wife's license plate over 700 times in three months, citing "test" as the reason.
• Illinois police officers conducting searches for federal immigration agents, using "immigration" as the justification.
• A Kansas police chief searching his ex-girlfriend's plates 164 times and her new boyfriend's 64 times, presumably not for an Amber Alert.

Growing Backlash and Counter-Surveillance Efforts

The widespread use and potential for abuse of Flock cameras have led to a growing backlash. Cities like Austin and Evanston have voted to remove their cameras. A lawsuit in Virginia argues that this system constitutes an illegal dragnet that the Fourth Amendment was designed to prevent.

Despite these challenges, Flock Safety continues to expand its offerings, including:

• Gunshot-detecting microphones
• Police drones
• An entire law enforcement operating system called Flock OS

The company also had a partnership with Amazon Ring to allow police to request doorbell footage through Flock's platform. However, this partnership was terminated after public outcry following a controversial Super Bowl advertisement by Ring.

In response to the proliferation of Flock cameras, a project called DFlock has emerged. This open-source initiative, started by software engineer Will Freeman, aims to map the locations of Flock cameras across the country. Freeman began the project after noticing cameras in Alabama and finding a lack of information about their placement and approval. He built a map using OpenStreetMap, and the open-source community has contributed to its expansion. Flock Safety's lawyers have issued a cease and desist letter to Freeman, which he has reportedly disregarded.