You know the scene: you're driving along when suddenly a glint of reflected light from a dinner-plate sized solar panel attached to a nearby lamppost catches your eye. You're reasonably sure that panel wasn't there last week, yet here it is now. Attached to the pole beneath it is something that looks like a hand-crank pencil sharpener, only plastic. You have just seen a Flock camera.
For today's lucky 25.000, Flock is an AI-based automatic number-plate reader, owned and operated by the dubiously-named company, Flock Safety. For about the last 3 years, Flock has been gradually taking over the United States' closed-circuit traffic enforcement cameras, allowing police departments and intelligence agencies all over the country access to millions of records all at the same time. What once required a search warrant, signed by a judge, now requires nothing more than logging into your Flock account and clicking a few things. Advocacy groups have called this a blatant disregard of rights afforded people under the Fourth Amendment (protection against unreasonable search and seizure), which it is, but as with everything else in the late-stage corporate feudal state, it's more insidious than that.
The language of Flock Safety's contract that city councils all have to sign in order to finalise the deal and authorise the installation of the camera equipment is extremely sketchy, leaving Flock Safety as the sole owner of all data collected by their camera network. Also, it holds that all previous contracts are fungible and can be changed by the company without prior notice, authorisation, or the ability to legally object. Especially as the Trump regime is heavily invested in AI technology to provide a focus for ICE operations and National Guard deployments, Flock Safety is able to, in effect, hold a city hostage. Indeed, there are 4 documented instances of Flock cameras continuing to be used after the city's contract expired or was broken-- namely, Verona, Wisconsin; Littleton, Massachusetts; Cleveland, Ohio; and Roanoke, Virginia. The company's CEO, of course, has an answer for everything. "Oh, we have a specific off-boarding period for cameras. Oh, the police department had an ongoing investigation and requested extended access. Oh, we were never told that the city council had backed out of the contract." Shit like that. Furthermore, it recently came out that the Department of Homeland Security is authorising the use of Flock cameras without consulting the city council, the state legislature, or really anyone in a decision-making position for a particular municipality. Totally co-incidentally I'm sure, this tends to happen more frequently in cities with liberal mayors or a predominantly liberal city council.
Recently, Flock Safety CEO, Garrett Langley, announced the company was addressing several issues related to abuse of the system by individual police officers; such as implementing an auditing tool that is purported to flag inappropriate use of the system and lock that user out. Also, users must provide a code that matches the one in the record before a search can be run, and all searches will be logged. The company said that this one in particular was something that advocacy groups had been hounding them about.
In the words of Joseph Sisko, "There isn't a test that's been created that a smart man can't find his way around." In other words, the new "accountability measures" are just a palliative smokescreen to get the Senate off its back. The AI-powered auditing tool will work just about as well as you could expect, and the individual cops who abuse Flock systems to stalk their exes and keep track of their wives will continue being able to do so. After these "accountability features" were announced, all the advocacy groups who were, theoretically, interested in maintaining the right to privacy put their thumbs up and went, "WORKS FOR ME!!"
However, we still have not addressed the elephant in the room that is unconvincingly disguised as a grand piano: Flock cameras are networked. All the video captured, all the analyses made, all the data collected goes into a central datacentre, owned by Flock Safety. The computers are theirs, and according to the contract for services, the data is theirs, too. Therefore, it is at Flock Safety's sole discretion what is done with the data that is collected.
I've been trying not to bring this up for years now, but I can't not talk about it anymore. Data correlation.
What does that mean, then? Networks of any kind are basically datapoints for a central bank of computers. In order for a network to function as intended, the host needs to know which clients are logged on at any given time. If the network doesn't know that client #64 is logged on, then it won't know to serve information to it or accept information from it. However, starting in around about 2009, monolithic software companies discovered how to track individual peers across multiple different networks and began to spin off statistical aggregators to blaze new trails into advertising. Over time, more specialised data could be collected through the use of smartphones and home assistants, as well as consolidation of resources through mergers and acquisitions. Before there was ever such a thing as "Flock Safety", Google, Amazon, and Facebook were harvesting user data in this manner to sell to advertisers. A friend of mine in college noticed that after she searched on Walmart.com for swimsuits, her sidebar on Facebook began to fill up with adverts about tropical vacations, swimming pool installation, and the Top 10 Ways to Prevent Sunburn This Summer. In a somewhat less amusing tale, someone I follow on Tumblr talked about how, after she used her smartphone to call her brother and discuss whether it was time to put their mum into assisted living, her adverts on Instagram suddenly all had to do with funeral homes, burial services, and bereavement counselling.
These examples are very simple demonstrations of data correlation using a single datapoint. What happens, do you suppose, when these same advertising systems are given data from multiple points? Do you suppose it goes, "Too much information, does not compute"? Consider that the very first computers, back in the 1940s, '50s, and '60s were being used to analyse relatively large amounts of data and provide a single response. They were doing data analysis back in the literal 1950s, using mere kilobytes of memory space. In 2026, consumer-grade desktop PCs routinely have 4,000 gigabytes of physical storage and 32 gigabytes of memory. The entire computation history of the ENIAC would take up less space than a single photo taken with any model iPhone. Data analysis and correlation is no problem for a modern computer and, with all the physical storage available to them, Google, Facebook, Amazon have the ability to analyse and correlate data from literal billions of peers.
I say that I stay masked so long after COVID (which never really went anywhere, but let's not change subjects just now) because I don't want the corporate feudal state knowing what I look like, that I want to stay a question mark in its computers and a black mark on its record. However, I know that, using its vast dragnet of doorbell cameras, individual security systems linked to cloud storage, home assistants, and even my own phones, I am just as much of an open book as anyone else. I haven't seen an advert on any website since I switched over to Firefox full-time in 2014-- it doesn't matter. Even though the targeted advertising can't reach me directly, it does not preclude the systems' ability to compile a jaw-droppingly accurate profile on me to whoever can pay. You see, I'm the same as you in this regard: everyday, I use the same things in order to do things out in public. I use a smartphone, an internet-connected computer, a car, a credit card, the roadway, service providers, and stores. Also, my state's DMV made a big deal about using everyone's ID photos to train AI facial recognition systems back in '22.
Let's go through a typical day, shall we? Every afternoon, my neighbour's Amazon Ring doorbell camera sees me and Riza go down the stairs. We've done this enough for Amazon to know what our walking gaits look like, what we have in our wardrobes, and potentially whether we have done the washing recently or not. It also sees when we carry goods up to our flat, identifying which stores we go to. If something is too large to fit in a bag, Amazon can see what that product is. It can also see when we take the empty box that the product used to be in out to the skips. More to the point, Riza and I encounter this camera shortly after we leave the house, therefore it knows how we are dressed and what accessories we tend to wear the most often. Since Amazon currently has thrown the widest net over the world, other cameras across town that feed into Amazon Web Services can correlate our position and determine our most likely destination based upon factors like known work schedule (Riza's employer uses AWS to keep track of their employee records), habitual behaviour, frequented stores, and frequently-purchased products.
In sections of town where Amazon's net has not been effectively cast, Google steps in to fill the informational void, using our Google Advertising IDs broadcast by my phone and Riza's tablet to determine where we are and whether we are contributing to a traffic backup. Thanks to Android, Google has the most specialised knowledge of our lives. Our Android devices are always in the room with us, and we know from the 2013 PRISM dataleak that microphones are always open, cameras are always running, data is always cycling in and out of all Android devices all the time. Google has heard our conversations, our inside jokes, when we cry, when we laugh, when we have sex. It knows that we're transfem, we're lesbians, we're engaged to each other, that I'm unemployed, about this website, all the passwords I've ever entered either on the phone directly or within earshot of the acoustic keylogger, the contents of all the documents I've ever written with this keyboard, the fact that I'm a media preservationist, what media I've preserved (because I tend to talk about it fairly frequently)...
Suffice to say, the list is extensive.
Facebook cannot analyse me anymore, however, because I've been away from there since 2013 and I have no designs on going back. There may be an advertising partner or two hundred that Facebook is continuing to feed outdated information to for a swift quid, but my immediate concerns are Amazon and Google.
Fortunately, Flock Safety has not made significant inroads into our city council here. While a few large shopping centres use Flock instead of our phone company's cloud camera services, the roads are not generally being monitored. A few are, but not enough to cast a dragnet. However, some cities use Flock religiously, and that's where the trouble is. Amazon's AI is generalised in the way it makes observations about people. Flock's AI is specifically designed to make note of identifying characteristics about cars. Make, model, year, bumper stickers, window decals, dents, chipping paint, and certainly number plates. The efficacy of Flock's AI in making an accurate identification is dubious at best, but even as granular as it can be, it still can be used as another link in the chain to personally identify you and determine where you go and what you do every minute of every day.
The sad truth of the matter is, even if Flock Safety went out of business tomorrow, it isn't a major player in corporate feudal infrastructure. Even before Flock showed up, police agencies were using Amazon's network of doorbell and cloud cameras to track people's activities and Google Advertising IDs to make targeted arrests. Basically, Flock is only significant because of the CEO's close association with the Trump regime. Garrett Langley went on record, saying that people who interfere with Flock cameras should be considered domestic terrorists. Flock Safety routinely provides data to ICE, Border Patrol, the FBI, and the DHS, and ICE has used Flock data to make arrests. The specific trouble with Flock, relative to Google, Facebook, and Amazon, is that Flock's CEO is a bootlicker. The corporate trifecta has enough money to sway court decisions in their favour and even sue the regime, and Trump knows this. He's been trying to strongarm Jeffy, Sundar, and Mark for most of 10 years now, but he always has to do it from a distance because of how influentially wealthy these boys are. Garrett Langley, on the other hand, has an up-and-coming data collection that he is perfectly willing to share with the regime. While Amazon would probably charge the Justice Department for access to its files, Flock just hands them over, sometimes without any prompting.
So, Flock may not be the most ubiquitous data collector in the US, but it certainly is the most dangerous.