Here's The Unimaginably Huge Number Of Pictures Taken By Flock Cameras Every Month - Jalopnik

Joe Raedle/Getty Images

Chances are, you strongly dislike the Flock cameras that have cropped up in your city or town, and all across the country. Thanks to them, it's certifiably no longer a stretch to say that Big Brother is watching in America, but how often are these cameras watching you, exactly?

According to Flock, in a blog post from February 2026, "The network processes more than 20 billion vehicle reads each month across 49 states." For simplicity's sake, let's take 20 billion photographic plate-reads at face value, from just a single company (that doesn't include the cameras from companies like Axon that are replacing Flock in some locations). Some back-of-the-napkin math gets you to the point where it's a little over 408 million reads per state, per month. And, keep in mind, that was in February, and we wouldn't be at all surprised to learn that the number has grown since then. 

Assuming the month you're talking about here has 30 days in it, the cameras are taking photos 13.6 million times a day in each of the states they operate in. Divide that by the number of seconds in a day (86,400), and that's 157 reads per second, per state on average. Go ahead, count to three real quick. Almost 500 people's license plates had their picture taken essentially without consent in just your state in that tiny amount of time.

How many of those reads are wrong?

Justin Sullivan/Getty Images

The even more problematic implication of that astronomical number, of course, is the system's accuracy or lack thereof. Per Flock, the ALPR network they have in place reads license plate right 96% of the time, which feels high on the surface until you realize that, given 20 billion monthly plate reads, that 4% error rate equates to 800 million potential nonsense traffic tickets, wrongful stolen car arrests, and, in the worst case, dangerous and unnecessary police encounters every single month.

Again, that's assuming you take Flock's own accuracy claim as gospel. The town of Roseville, California, which you'd probably never heard of before last month, made headlines after it found its Flock camera system was only accurate 29% of the time. No, we didn't get our numbers flipped — it was right just 29% of the time, which, said another way, is that it had a 71% failure rate. Maybe Roseville is an outlier; maybe it isn't. But for illustration's sake, if Flock cameras as a whole only had a positive ID rate of 29%, that 800-million best-case f-up number becomes, wait for it, 14.2 billion potential times law enforcement might stop someone over a technical error.