What happens when you drive past an ALPR?
An automated license plate reader can create a searchable record tied to a plate or vehicle description, together with the time and camera location. Depending on the product, the record may also include visible characteristics such as vehicle type, color, damage, accessories or direction of travel.
One read is a point. Many reads across a network can reveal a route, routine or pattern.
Pass the camera → create a time-and-place record → store it → search it later → connect it with other observations
How much can these systems actually see? →
Who can access the data, and how long can it remain? →
Why is that different from an ordinary camera?
An ordinary security camera generally records a scene for later human review. It can still raise privacy concerns, and some modern video systems add automated identification or search.
An ALPR network is purpose-built to extract and index vehicle observations so an authorized user can search them later. Depending on permissions, those records can also participate in interagency or wider networks.
| Ordinary scene camera | Searchable ALPR network | Why the difference matters |
|---|---|---|
| Records video of a place. | Extracts vehicle or plate information with time and location. | Indexed information is designed for rapid later search. |
| Often reviewed manually after an event. | Can search by plate and, depending on the system, vehicle characteristics or connected data. | Searching at scale can reveal movement patterns that one clip cannot. |
| May operate as an isolated system. | May participate in agency-sharing or network access. | The practical reach can extend beyond the cameras a city owns. |
Not every roadside camera is an ALPR, and not every ALPR deployment has every available feature. We verify the product and local records before making a local claim.
Can these systems help investigations?
Yes. ALPRs can help locate wanted or stolen vehicles and develop investigative leads. We do not dispute that.
Investigative usefulness is evidence to consider. It is not, by itself, a limiting principle. More surveillance can usually produce more information; a community still has to decide whether that benefit justifies routinely creating searchable location records for the much larger number of people who are not suspected of a crime.
What effectiveness research actually shows →
Public safety without mass tracking →
Why care if you have nothing to hide?
Privacy is not the same as secrecy.
We close restroom doors, draw curtains, use passwords and choose who receives personal information. Most people do not publish a minute-by-minute log of everywhere they go. None of that implies wrongdoing.
Being visible at one moment in public is different from having many observations automatically indexed so they can be searched, shared and combined later.
Why ordinary privacy still matters →
What can go wrong?
In February 2026, reporting documented a Flock plate misread in Sherwood, Arkansas that contributed to police conducting a high-risk stop involving an innocent couple and their infant before the mismatch was discovered.
In Alpharetta, Georgia, internal-investigation records reviewed by WIRED showed an officer repeatedly used Flock searches involving a former romantic partner and another officer. The officer resigned after the investigation; WIRED reported that a criminal investigation was continuing.
The first example involves machine output and human verification. The second involves misuse of authorized access. Neither proves that every read is wrong or every officer misuses the system. Both demonstrate risks a local government must evaluate before creating the capability.
Documented misuse, failures and harm →
Flock claims compared with the public record →
What about safeguards?
Shorter retention, individual accounts, required search reasons, audit logs, restricted sharing and warrants where required can reduce certain harms while a system exists.
They do not answer the threshold question: should government routinely collect these observations about ordinary people in the first place?
Always separate four layers:
- Vendor capability — what the company says the system can do.
- Local configuration — what this agency actually purchased or enabled.
- Written policy — what local rules say users may do.
- Actual use — what audits, searches, incidents and other records show happened.
A promise at one layer is not proof of the other three.
Why removal instead of better rules?
We oppose routine, searchable collection of vehicle-location data about people who are not suspected of wrongdoing. Safeguards can reduce some harms, but our goal is nonrenewal and removal of the mass-tracking infrastructure.
That is a policy conclusion, not a claim that every user acts badly or every camera reading is wrong.
Who decides?
Local residents and elected officials do.
DeFlock Metro Detroit is an independent local civic group. We obtain records, explain the technology, document risks and help residents participate before surveillance contracts quietly renew. Local control includes the right to support the cameras. It also includes the right to organize against them.