Flock Cameras Are Not Even the Worst Part of the System

Now, I'm happy to see that Flock cameras are back in the news because they are newsworthy. And I'm glad to see that most of the opinions that I see being published about them by the general public are negative because the ways that Flock cameras can be misused are pretty negative. However, even though I don't want to take anything away from the cautions that we have about the cameras, the cameras are only the tip of the iceberg. And so today I want to talk about what the real threat is inside of Flock Safety, the company, but also similar companies that are coming and rather inevitably. [music] So, this is a Flock camera. And not only is Flock Safety, the company, relatively inevitable, this camera type was inevitable. A remotely operated camera that has a little LTE modem on the side that lets it talk to cell phone networks and communicate with the cloud. They generally all at this point have solar panels on them, makes them pretty easy to spot. But ultimately, the hardware of these different cameras is not that different from your phone. It's got a camera, it's got a brain, it's got storage, it's got a battery. The solar panel is the only kind of new thing and it talks to the internet and puts its stuff in the cloud. As these components got cheaper and more ubiquitous, these solar-powered remote cameras with no connection to the internet requiring infrastructure and wires, just cell phone towers off to the side somewhere, were inevitable. And cameras that wanted to build these things and offer them as services were inevitable as well. That's why Flock Safety, the company, is the only person getting a lot of attention for cameras like this right now, but they're not the only player. But the issue with these cameras, are many. The hardware of the camera means that the storage device inside of here that captures all the imagery is not secure. Physical access to this device is simple, and then the data that is on there is accessible. Then, the way that that data is transmitted to the cell phone networks, also not entirely secure. But the biggest issue that I want us to focus on is the giant database full of information up in the cloud, where all of the data captured by the camera lives after it's transmitted from the cell phone network. This right here is where we want to spend a lot more of our attention. I would say a lot more of our negative attention. This is where the misuse happens. This is what separates Flock safety cameras from regular security cameras. This is where we need to be attacking the misuse of the data and the large database and the AI models that manipulates the data inside. If we get too focused on the cameras, I think that we miss this. Now, the cameras are easy to focus on because they have many problems. And they're the physical manifestation of Flock safety in our communities. They're mounted on this pole right here, which is very fragile when you want to cut it down and very not fragile when you crash into it with your car. Just the installation of these cameras is a problem on top of everything else. And as these cameras have developed, they have gotten new features. They have not just been taking pictures of license plates, they can now track video. They now have different sensors on them, kind of like your phone, to detect Bluetooth and wireless devices like your phone going by. So, now you have the ability to collect all kinds of data and put new kinds of data inside of the database and then using fancy AI models connect some of these things together and make inferences about them which we'll get to in a minute because the RF data collection here is just one of the many new tricks that Flock is bringing to the table. They have drones now. The main goal is to have a multi sensor database and a fusion system that ties all of those things together. [gasps] So I guess the question is why why is that a problem? Well, there's a bunch of different reasons. We'll get into some of those in a second, but I want you to think about all the data that can be collected by this camera right now. Originally the pitch was these were license plate readers and so as um we have our database here and our database grows to where it contains more and more information. It contains information not just on license plates captured, but it begins to build out a map of say different areas where traffic is occurring and by putting a Flock camera here and a Flock camera here, you can grab all of the license plates, all the traffic that moves around. Once you do that, you can now assume the paths that cars were taken based on when their license plates were detected on that particular camera and then that goes into the database not just as a time and a place, but you start to have database entries that contain vehicle locations, but also their routes and some of that has to be estimated cuz there's no camera right here. You can see where that starts to be an issue. But as the cameras got more sophisticated, they started to collect other things about vehicles, not just license plates, but color vehicle make. These are all things that Flock has advertised, so we know these things are relatively relatively understandable. Occasionally Flock will announce that they've used these cameras to help find missing people that are not in vehicles. And so, we know that they're capturing other stuff as well. The opportunity to capture face recognition and data, connect it to license plate data, connect it to other stuff, that's huge. If these things have significant video capture capability, let's say up to like 15 frames a second is enough to make relatively good guesses about people's skeletal structure and gait, the way that they move and walk around is a thing that you can extract from raw video now using AI models, and you can tie that to a person or to a license plate on a car. Then, of course, the person is carrying with them a device, most likely probably a phone. Maybe they also have a smartwatch, and maybe they also have earbuds. All of those things emit signals that have unique identifiers that you can connect to this device. It will see those things passing by, collect that data, send it to the cloud, and to this large database. And again, fuse all of these different things together. And then the car itself also has a number of other things. It's not just the license plate, but your vehicle almost certainly, if it's anywhere near modern, has sensors inside of each of the tires, tire pressure monitoring system. That's a wireless system that emits unique IDs that can be picked up really easily. I was doing it a couple of weeks ago for reasons that I will go into in future videos. But all of this data can be collected by the existing hardware very simply and very easily. And all that data can be collected and put up here and fused together. This database is the main Flock product. The Flock operating system is what is being sold, not just the cameras. So, yes, focus on the cameras and all that they imply, but don't lose sight of this larger tool that is getting built. Uh the other thing I want to point out is as there is more and more negative attention on these cameras, there's a greater opportunity for Flock to use these temporarily. My personal theory is that these cameras only need to be temporary. I mean, the ones that are installed now are temporary. The battery that they have inside of them out there in the hot sun is only going to last for a couple of years. And the more of these things they can sell, uh the more cash Flock is able to raise, but over time they will very likely replace these devices with just RF detection. There are Flock cameras out now that have a little radar module on the bottom. We're not sure exactly what it does cuz it doesn't have an FCC ID on it that we can look up. Um but that is collecting RF data in some fashion. And one of the things that Flock could do as negative attention grows, they could admit that they have heard the general public and they have gotten rid of the cameras. After all, cameras do have weaknesses, not just privacy weaknesses, but um they have to see you, which means they have to be in places where you can see them. And so, it's really easy to pick these things out. But once you replace the camera bit with just RF detection sensors and radar, well then you can put these things in more places and you can detect more stuff in a way where the device itself is much less detectable. And if Flock has many, many years of video training data alongside all of that RF data and they've already built the models to connect these things together, they can and probably will make the argument that the camera data is no longer necessary, which is theoretically a win for privacy, right? Cuz you're no longer having your image collected. But all of your data, which is theoretically all of your movements and all of your activity, that is still being collected. And there's more data that is out there to collect. Obviously, if you have a sensor package that can collect all this data, then you can load it into your database. Uh but even without these devices, you know a lot of this data is being collected. Your phone, um well, it's uh it's it's a surveillance device like like, you know, like the Flock camera. And it is connected to these cell phone devices all the time. Your earbuds and your watch probably aren't unless you have one of the newer watches, but if you connect those to this device, all that data can go into the cloud. Modern cars have uh 4G or LTE modems that transmit telemetry data up here, and that goes to mostly car manufacturers who sell it to insurance companies. But uh my theory is that the bigger this database is, the Flock safety builds for law enforcement purposes, um the better pitch that they have to the federal government, other law enforcement agencies, regulatory and legislative groups, that since they have the biggest database, they should own some of this other data as well. They should have all of the mobile data that are collected by the different carriers, all the location data. They should have that in the big database cuz the database is more useful the bigger it gets. All of this car information that is tracked by the different manufacturers, well, that should live in the big database cuz the bigger the database is, the better. So, a lot of what Flock is doing right now with their cameras, I think is building a tool that really has nothing to do with cameras at the end of the day, and is a chance to grab market share and make a bid for bigger data. And ultimately, the data is going to be the product, not the images. The images are just the way that we collect the data at the moment. Now, uh there are other companies trying to get into this market share. There is Axon, they make body cameras. Now, they are the body camera manufacturer in the United States. Most of the body cameras used by law enforcement are Axon cameras. There's some backstory there that's kind of interesting, but grabbing some of this market share so that you are the monopoly makes sense when you're in this space. There's a lot of areas where you're just selling a product and any market share is valuable. But when you're going for being the sole provider of the biggest database possible, having the best AI models because you have the best training data possible, you really want to be the winner there. It's There's not a lot of room for second place, third place, or fourth place. And they talk a lot about pattern of life. There's a huge value in developing a database where every person is connected to devices and movements and they develop a pattern of life for this person and for this person and for these people over here. And they pitch this as a crime-fighting tool. But there are problems with this. If your goal is to build a giant huge database that just has really good, clean, orderly data where patterns of life are clean and predictable and your inferences are good and solid because the data is nicely joined and all of the fusion between units works well, then the outliers become a bit of a problem. And the outliers are where a lot of your criminal activity exists. So this pattern of life that gets developed and these movements that can be adequately tracked and well guessed, those work well for people whose lives follow patterns. And a lot of the people that we want to do the, you know, crime-fighting activities against don't have these same patterns. You know, people who steal a phone one day and trade it for meth the next day and exchange for the use of a car that's at a flop house and they're only in it for maybe 12 hours before something else is going on. These guys don't have a pattern of life that fits in this area. Now, an argument could be made that looking for people that don't fit into the pattern is good police work, but I don't anticipate that is what is going to happen. And the reason for that is the way that Flock has handled some of their other outliers. So, one of the big um pervasive scandals that Flock Safety has had to deal with is misuse of this data. Because part of the issue is this data, the big database, is from many different jurisdictions. And Flock puts cameras and data access into private hands. Uh Home Depot, for example, has lots of Flock cameras in their parking lots and they have access to that data. They have access to this subset of the data because it's their property. However, this Home Depot data is also accessible to everybody in law enforcement with the appropriate permissions. And Flock is pretty liberal uh and lavish with the permissions that they give. And so, there have been many cases, many many many many many cases where officers uh of the peace are using Flock access to do things not related to crime-fighting activities. Stalking of people, tracking of people, unsavory sort of things. If YouTube suggests you other Flock camera content uh in your sidebar, you'll probably find a bunch of these stories. And the way that Flock has decided to combat this is to no longer save the user ID that requests data. That way people can no longer get caught doing this sort of thing. No longer get caught doing this sort of thing. Uh but even with the information that we do have, we do know that lookups of people um and license plates and traffic, and just various pattern of life stuff happens predominantly outside of specific case files. So, in some jurisdictions, it's it's about 80% to 85% of all queries of the data. There is no case ID. There is no reason given for the search. The vast majority of what's being collected and what's being asked for does not appear to be directly related to catching specific criminals. So, yeah, as as this develops and pattern of life becomes a product that is sold to law enforcement agencies or to municipalities or to other types of bureaucrats that are not actually the ones working on the street, my expectation is the focus will be on data cleanliness as opposed to the outliers that law enforcement actually should be focusing on. Now, at this point, I would imagine that in the comment section, everybody is annoyed or hopping mad cuz that's the YouTube comment section. And I bet that I've annoyed people in such a way that half of the comments are annoyed with me that I am soft selling the problems with Flock and that I haven't called for the guillotining of everybody with a badge. And then on the other side, there's some people that feel the exact opposite. This is a great crime-fighting tool. This will be used to deal with the obviously real problems with crime that we have in the United States. Why not let people have this equipment? And the answer to that is that I don't believe that this actually helps the real job of law enforcement. The real job of the individual law enforcement officers who are attempting to bring justice to actual victims that have had actual crimes committed against them by actual criminals. Once you have a tool like this, it turns into a surveillance and supervision tool. There's almost no way to build something like this that is all about a giant database, making sure that it has the maximum amount of data possible, that it has the maximum number of people inside, the largest possible database of faces, the largest possible understanding of how people travel around a city, and the patterns of life that fit inside super communities without it focusing on collective actions rather than specific crimes and facts. And if it's true that our negative attention brought against the flood hammers themselves result in no more images being captured while we still collect everything else, all the other metadata. In some ways, that's worse. If I'm ever accused of a crime because my vehicle was nearby or it was estimated to have gone in a place that puts me near the scene of a crime and my other devices are somehow sort of in the vicinity, I would actually love for there to be exculpatory evidence from cameras of me not being in the car or me being under duress or being threatened or kidnapped by a certain person. It's very possible that the actual factual evidence collected by cameras would be far more helpful to me than this generic system here, especially in the age of AI when this generic system here doesn't exist just as a query system for random records of person A or B or person 23879-whatever. This becomes a large complicated network of nodes and inferences in an AI model where all of the activities of these people with predictable patterns of life are used to predict other people. If certain activities that you participate in line up with certain activities that the experts have decided are problematic, well, obviously you see where that might potentially go. You're building police work over estimates and guesses and statistical probabilities instead of hard facts, which like it or not, cameras could actually pick up on. Again, I'm not defending the cameras themselves. I'm saying that things could get bad enough that we will wish that we had them back. And I have an example that I think would be helpful. These tools could easily be misused in the way that London's closed-circuit television cameras could be [music] misused. For most of my life, London has had more cameras per capita than any other city. It's the most surveilled state when it comes to just cameras themselves. And over the years they've built different systems for integrating those cameras. I don't fully understand where it's been at all the different times of that tool or that bad measures lifetime. But I do know that for the last couple of decades, London has had a problem with rape gangs. There have been inquiries published that really indicate that this is a systemic problem. And interestingly, that massive network of cameras that are still cameras that do still locate individuals, that do record data that can be looked at by human cops with their actual eyeballs. Those constables that were not on patrol and instead looking at screens information gathered by cameras were not able to find or bring to justice really any of the rape gangs found in the inquiries. But they were very very capable at finding all the people that attended protests against the rape gangs. So, that is the kind of misuse of a system that I want us to be thinking about. And it goes far far beyond the cameras. The law enforcement officers that I talked to definitely want more tools in their tool bags. That's part of the reason that T-Rex exists is to make tools for people that need to carry life saving equipment. And that does include magistrates as well as regular citizens. Now, [music] T-Rex still has the policy that we will not sell anything to law enforcement officers in jurisdictions where the private citizen cannot own the gear. But nevertheless, a civil magistrate is a job with a responsibility. He carries the sword and he doesn't carry it for nothing. The problem is that a lot of our civil magistrates don't need tools that help them find aberrant patterns of life. They don't even need tools that help them find bad guys. A lot of the times they don't even need tools that help them catch bad guys. If you read in the papers, you will find time after time after time horrific crimes being committed by people that have been arrested 20, 30, 40 times and let go way way too often. This is not a system that solves that problem. This [music] is not a tool that addresses the actual root issues that we have in our communities. Let me check my notes here and make sure that I covered everything that we wanted to, all the potential misuses and problems. Yeah, I know, we just barely scratched the surface. I will not be pushed, filed, stamped, indexed, briefed, debriefed, or numbered. I am not a number. I am a free man.