Flock cameras have become a common sight in American cities, yet discussions about privacy and surveillance often overlook the vast scale of their operations. The company’s automated license plate reader (ALPR) network handles over 20 billion vehicle reads monthly across 49 states, according to a February 2026 company update. This excludes competitors like Axon, which have expanded into some regions since then. The figures translate to roughly 408 million reads per state per month, or about 13.6 million daily, assuming a 30-day month. Breaking it down further reveals 157 license plate captures every second, meaning nearly 500 plates are photographed without consent in a single state within just three seconds.
The sheer volume raises concerns about misidentification rates. Flock claims its system achieves a 96% accuracy rate, but applying this to the total volume reveals a troubling reality. A 4% error rate across 20 billion monthly reads equals roughly 800 million potential false positives each month. These could result in wrongful traffic stops, mistaken alerts for stolen vehicles, or even dangerous police encounters caused by technical failures.
The risk is not just theoretical: in Roseville, California, officials found their Flock system had only a 29% accuracy rate, meaning it failed 71% of the time. If this were the national standard, the 800 million monthly errors would surge to 14.2 billion, turning isolated mistakes into a widespread problem.
Read Also: Hyundai Grandeur Stuns Looks Falters on Road
Flock’s own data hints at recognition of the issue. The company’s blog post acknowledges that false positives can lead to “unnecessary police stops,” though it does not explain how errors are addressed. The 157 reads per second figure highlights how rapidly surveillance systems operate at scale, often without public scrutiny. Even if most captures are correct, the potential for errors becomes a civil liberties issue when scaled across millions of interactions.
These numbers also ignore other critical factors. Some states enforce stricter privacy laws limiting how long agencies can store plate data, while others permit indefinite retention. Flock’s system does not differentiate between a stolen car and a misread plate—both could trigger alerts. While the company’s accuracy claims may hold in controlled settings, real-world conditions like weather, lighting, or damaged plates can further reduce performance. The 20 billion monthly reads represent more than just an operational milestone; they reflect a surveillance infrastructure functioning at a pace most people cannot fully grasp—one where mistakes are not rare but inevitable at this scale.
The lack of public transparency worsens the problem. Flock’s blog post does not reveal which states experience the highest error rates or how often false positives lead to enforcement actions. The 14.2 billion potential errors in a worst-case scenario are not mere speculation, they illustrate how quickly a system intended for public safety can become a significant liability when deployed on this scale. The core issue is no longer whether these systems function, but whether their advantages justify the risks when expanded to this extent.
