Flock Safety looked at the cars driven by Uber, Lyft, and delivery workers and saw a police camera network waiting to happen.
404 Media reported that a Flock presentation prepared for the Georgia Office of the Attorney General said a proposed partnership with dashcam maker Nexar included “350K Uber/Lyft and other delivery service devices”. The plan would have used cameras in privately operated vehicles to collect license plate data as drivers moved through their normal routes.
Flock told 404 Media that the partnership was never executed. The report found no evidence that Uber, Lyft, Nexar, or individual drivers agreed to participate. Neither rideshare company nor Nexar responded to the outlet’s questions.
Flock says the idea went no further, and the available reporting backs that boundary. The sales slide still exposes what the company believed it could build: cameras carried by working people, feeding a surveillance network used by police.
A dragnet that moves
Flock’s existing license plate readers capture plate images, vehicle characteristics, timestamps, and camera locations. The company says most communities keep the data for 30 days, every search is logged, and local customers control sharing.
A fixed camera watches one road. A mobile network could follow the paths that workers take all day, filling the spaces between poles and reaching residential streets, parking lots, medical offices, houses of worship, and protests. Flock’s presentation put the proposed scale at 350,000 devices. The slide is a sales claim rather than a verified deployment count. It still tells us what Flock thought was worth selling.
Nexar’s own privacy policy says its cameras can generate images, video, sensor readings, and location data. It specifically includes rideshare drivers and fleet managers among its users. The policy also says raw images can contain people and license plates, while some data used for mapping is anonymized on the device.
The proposed technical design remains undisclosed. The record does not say whether Flock expected raw footage, plate records, vehicle descriptions, or a filtered stream. It also leaves open whether a driver could pause collection, who would pay for the data connection, and what a passenger would be told. Those unanswered questions should stop any procurement. A surveillance vendor does not get to answer them in private.
Local cameras already became a national system
Flock presents its product as a narrow investigative tool. Its Trust Center says the system is for authorized public safety cases and prohibits mass surveillance or tracking people.
The network’s real scale makes that promise hard to accept. The Associated Press reported last year that Flock cameras were mounted in more than 4,000 communities and captured billions of plate photos each month. A local agency may own its data, but other agencies can ask to search it. A system can be locally purchased and nationally useful to police at the same time.
Flock’s safeguards have already failed in cases that mattered. The Electronic Frontier Foundation obtained search logs and court records showing that a Texas sheriff’s office queried more than 83,000 Flock cameras in an abortion investigation. The search reason entered into the log said, “had an abortion, search for female.” Flock and the sheriff had publicly described the case as a welfare search, but the records identified it as a death investigation involving an abortion the woman managed herself.
Illinois found another failure. The secretary of state’s office said an audit discovered that Flock had allowed Customs and Border Protection to access Illinois plate data, contrary to a state restriction on sharing for immigration investigations or lawful health care. Flock shut off the access and paused a federal pilot. In a company statement, chief executive Garrett Langley acknowledged that Flock had not created distinct permissions and protocols to ensure local compliance.
These were fixed cameras with public agency customers. Adding hundreds of thousands of moving cameras would multiply the collection points while making the system harder for any city council, auditor, or resident to see.
Keep gig workers out of the surveillance business
Responsibility for the documented proposal belongs to Flock. There is no evidence that Uber or Lyft approved it.
They should still reject the idea in public. A driver should never discover that a dashcam used for safety or insurance also fed a law enforcement database. Consent buried inside a camera policy or platform agreement would be worthless. The choice must be separate, informed, revocable, and free of any penalty to a worker’s access to jobs.
The same rule should apply to passengers and everyone outside the vehicle. A person driving to a clinic, attending a union meeting, visiting a mosque, or joining a protest does not lose privacy because a delivery driver passed by. Flock says its cameras capture vehicles rather than people. A plate linked to a time and place is still a record of where a person likely was.
Police have a real argument for mobile plate readers. A moving camera could help find a stolen car, a missing person, or a suspect in a serious crime where fixed cameras have blind spots. Flock says its searches are tied to accounts and recorded for review. Those uses can justify a targeted tool under strict rules. Bulk collection followed by later searches reverses the presumption that surveillance requires a reason.
The rules must arrive before the cameras
Virginia has already shown that lawmakers can set concrete boundaries. Its current license plate reader law requires most system data to be purged after 21 days, restricts outside sharing, mandates internal audits at least every 30 days, and requires annual public reporting.
A mobile system using privately operated cars needs stronger rules. Public agencies should need legislative approval and a public hearing before deployment. Drivers must opt in separately and be able to leave without losing work. Historical or pattern searches should require a warrant. Retention should be measured in days, sharing should be blocked by default, and independent audits should be public.
Flock’s Nexar pitch never became a product, which is good. The failed pitch left a warning: a surveillance company believed 350,000 cameras operated by workers could become another layer of its network and considered that vision presentable to a state attorney general.
The dashboard belongs to the worker, and privacy belongs to everyone traveling past it.