How should we think about automated license-plate readers, surveillance cameras, drones, and other public safety tools? Rafael Mangual sits down with Rahul Sidhu of Flock Safety to discuss the backlash against ALPR technology, the legal and privacy concerns surrounding surveillance, and the safeguards meant to prevent misuse. They also explore the real-world benefits of these tools in helping police solve crimes, recover stolen vehicles, and respond faster in emergencies.

Audio Transcript

---

Rafael Mangual: Hello and welcome to another episode of the City Journal Podcast. I am your host, Rafael Mangual, and I am so pleased to be joined today by Rahul Sidhu, who is an executive at Flock Safety, which is what we're going to be talking about today. Never in all of my years, Rahul, have I ever seen a backlash this intense. You know, the opponents of the use of what we call ALPR technology or automated license plate reading technology seem to be to me at least a little bit unhinged, but you know I like to try and take concerns seriously. So I think we can do that in today's discussion. And I just first want to start with a thank you for taking the time to come out and do this show. I think it's important, you know, for folks who are actually operating in the space to, you know, defend what they're doing and explain it to the public. So I just want to, you know, start off the conversation by commending you for that. Thanks.

Rahul Sidhu: Thanks for giving me the opportunity, man. Great to be here.

Rafael Mangual: Awesome, awesome. Well, I want to start with the basics, right? Because there may be some people who are watching and listening who don't really know what ALPR technology is or what Flock safety does. So why don't we start there, right? Talk to us a little bit about what ALPRs do, how long has that technology been in use? When did Flock first come on the scene?

Rahul Sidhu: Yeah, look, ALPR automated license plate readers, it's not a new technology, per se. I mean, this has been used for over a decade in policing. I mean, earlier than 2012 was ALPR piloted in the context of law enforcement. Flock itself was founded in 2017. It was not the first company to make automated license plate readers. You know, we, the company itself has been rolling out Flock cameras, which are just ALPR cameras made by Flock since then. Now, I think, I'll just like mention this right off the bat, that people will see ALPR, automated license plate reading camera, and say, it's much more than that. But like, let's just break this down really quick for what it actually is. It's a camera that is in a fixed position, typically like an intersection, a place in public, and as a vehicle passes it, it takes a photo of the vehicle. Now it derives, of course, it reads the plate, but it could also derive make, model of the vehicle, color of the vehicle. And that's where people say it's more than an automated license plate or your camera, which is reasonable. And then of course when you have a variety of these images from an intersection, you can say, it's taking a lot of photos. So that amount of data you know, is particularly useful for law enforcement. But again, this has been going along for quite some time, and that's what Flock's main product is. Flock also makes other products like live video, which is pretty similar to what you would buy from Best Buy. It's a live video streaming camera that goes directly back to typically law enforcement agencies, real-time crime centers, et cetera, placed in areas of high public safety concern. You might see them in areas with a lot of foot traffic, for example, or even intersections as well. Flock also makes software like our Flock OS software, which is essentially for these real-time crime centers to be able to look at different feeds and different information at the same time. Flock makes drones.

Rafael Mangual: Right.

Rahul Sidhu: the which is the company that I was CEO of before it was acquired by Flock, a company called Aerodome. we make our drones in Atlanta, Georgia. The drones, the drone stations, the cameras, et cetera, are made by Flock. those are drones that take off when a 911 call is you know made and they fly directly to the scene in  a 911 call and provide air support for the police and fire department as they arrive on scene. And then Flock also makes gunshot detection sensors similar to sh companies like ShotSpotter where if a gunshot goes off it can understand where the gunshot came from, date and time, and it's particularly helpful to curbing things like gun violence. So I'd say those are the products you know that Flock makes that people talk about quite often and often they actually confuse them.

Rafael Mangual: Yeah, I think that's I think that's right. And there are a couple of product lines that you mentioned that are interesting to me. I mean, you know, the stationary live surveillance cameras, you know it was one of them. I wrote a book a few years ago called Criminal Injustice, and I opened the book with a story that really kind of spurred the decision to write the book in the first place. It was a video that had gone viral for all the wrong reasons. it was a feed from a Chicago Police Department surveillance camera in a west side neighborhood called Austin, which is like, you know, typically one of the most dangerous neighborhoods in the whole country, but certainly within that city. And it depicted a murder. It was a a shooting, a drive-up shooting that claimed the life of a young woman named Britney Hill who was shot while shielding her one-year-old daughter who she was carrying at the time that the bullets rang out. And you know, it was such a visceral experience for me because I just watched this woman literally in the same act save her daughter's life, lose hers. I mean she l collapsed to the ground while this little girl was still clinging to her neck and bled out on a hot on hot pavement in the middle of the summer. you know, and the police were able to make an arrest very quickly in that case because it was caught on video, right? you know, they could see the car, they could see the the faces of the shooters, so they were apprehended within a few hours. And, you know, I wrote about that case because it turned out that one of the shooters had, you know, nine prior felony convictions, was out on parole despite having been previously convicted for things like second degree murder, you know, but that was one of those situations that stuck out that was like, you know, it's really good that they had this camera up there because who knows if i if they ever would have got that guy. And, you know, with a record like that and having done something like that, you know that if he wouldn't wasn't apprehended, that he would have, you know, continued to do that.

And you also mentioned gunshot detection technology, which I want to talk a little bit about because we just had actually the CEO of ShotSpot or or of Sound Thinking, Ralph Clark, on on the podcast a a few weeks ago, in part because you know his company I think was facing a very similar backlash to the one that that Flock Safety is is facing now. I mean, you know, the argument, as you know, is that these things are sort of being put in, you know, low-income minority neighborhoods and and you know, driving police to them and driving up racial disparities. And I think, you know, people who make that critique lose sight of the fact that there are benefits associated with you know having this technology in place, with getting help to shooting victims faster and getting them to the hospital quicker, which can literally save their lives. So yeah, I mean, how how does how does Flock sort of see itself within that broader community of tech companies providing this sort of stuff to police departments?

Rahul Sidhu: Look, I mean, I think that there's there's actually two things that I want to answer there. First is the question you just asked, and I also want to talk a little bit about the video cameras themselves. But to answer the question first, Flock sees itself as a company that's going to make technology for public safety agencies to improve the safety in their communities. That's the goal of it. Let's reduce crime, let's make people safer, let's go rescue missing children and missing people. That's the entire goal. All the products we make are designed for that particular purpose. There are a lot of ways that you can, you know, solve for improving, you know, creating benefits in the in the public safety ecosystem. We focus predominantly on those ways. Automated license plate readers, for example, I mean that's a main focus on if you have any vehicle related crime, things like stolen vehicles or someone kidnaps somebody, or just commits a crime and drives away. Like we're going to be able to provide both investigators an ability to solve that crime, more likely solve that crime, and an ability for a city to be able to ensure that every time someone drives a stolen car into that city to commit crimes, the police are notified and usually with a combination of drones, that vehicle is stopped. That itself has a pretty, you know, significant impact on crime. The other tools that we make for investigators and for real-time crime centers to be able to understand what's actually going on in their city to to make things safer. I mean, that's the whole reason we build these things. I'd also say that to to your point on the video cameras, this is something that police have been doing for decades. It's not new. The police have been putting these cameras in places of high concern for a very, very long time. Not even I'm not even talking about since the 90s. I'm talking about since these cameras, you know, existed even decades prior. And people don't really understand, like, why they're useful in some cases. I know I see the tweet constantly. People post a photo of a flock camera and they'll say, why is this flock camera reading license plates at this park? Or this like, you know, recreation area, the swimming pool, whatever the circumstances are. First of all, those are not license plate readers, those are regular streaming video cameras, they just look similar. But the reason you have them there is because, like, bad things happen in those areas. You can imagine if you have an area with children, that is actually going to be the highest likelihood area that you're going to see a stranger abduction occur. That's where the children are. So the police have a vested interest to protect those areas in real time, of course. But also from an investigative perspective, I was talking to a a command officer at an agency in Arkansas that I won't name because they haven't given me permission. But he was explaining to me that a few weeks ago, they have, you know, they've had a live, you know, Flock video camera in one of their parks for quite some time, for years. And just a few weeks ago, there was an instance where a man drugged and dragged a woman—this is at night when the park is supposed to be closed—behind a a restroom at a park and did not know there was a camera there and he you know sexually assaulted her. It was the most heinous crime he'd commit in that context. And they caught the entire thing on camera. And he left her there, and he, you know, went away. They obviously sent rescue after they realized what had happened. And because of the camera, they got the entire crime on tape and they were able to use it as evidence. Without that, there's a distinct possibility they would have never been able to do that. That man would not be convicted. He would be free to go do this to another woman, and that's terrible. That's what these things are for. It's to, you know, obviously aid in an investigation. Expect, you know, if you go to a park and you you hear, a man was abducted, you hear these stories all the time. Or sorry, a a child was abducted by a man in a park. Think to yourself like, okay, if we don't know where that child went, we just know the child wandered off and is now missing, that's terrible. You have no leads to go off of the likelihood of finding the child is much lower.

But if you have two things, one is a video camera that shows, the child was abducted by a man in a red shirt who got into a white Ford F-150 and drove off. And then you have a license plate reading camera at the nearby intersection that caught the F-150 and now actually has a license plate that they can tie to it, and then they can go, okay, I know who this vehicle belongs to. Your odds of finding that child and convicting that man for kidnapping are much higher. It's a fairly common-sense approach to law enforcement.

Rafael Mangual: Yeah. No, I think that's exactly right. I think that's exactly right. You know, we're kind of, I think, in the part of the conversation where we're talking a bit about the misconceptions and I want to sort of keep us there. But you know, I want to talk a little bit about the other capabilities that the ALPR cameras have, right? So they don't just read alphanumeric license plate. They can also, you know, sort of I think the term you guys use is fingerprint, a creative a vehicle fingerprint, right? So they can identify bumper stickers and damage and color or that sort of thing. You know, my understanding is that the benefits associated with adding those capabilities is really just responsive to the reality that criminals will often switch out license plates. And so the license plate alone, you know, doesn't necessarily you know, guarantee that you'll have a a better shot at finding the car that you're looking for. Was that the thinking behind adding those capabilities?

Rahul Sidhu: Yeah, I mean, absolutely. Let’s just like first isolate what the capabilities are. You're looking at vehicle characteristics. Like I said, white Toyota F-150 you know has a dent on it. These are unique very vehicle characteristics. The reason that the cameras are able to do that, and most ALPR cameras, not just Flock cameras, can do this now because this is what the police departments want, is because more often than not, a witness does not get a perfect license plate, full license plate, and says this is the vehicle that you know kidnapped that child. It's more often that they just catch a glimpse of like, I think it was a white truck. And they might notice something identifying. Like, I I saw, like, it had a, you know, a yellow bumper sticker on the back and it was a white truck, that's all I can tell you. They might get a partial plate, they say, I just saw the first three, you know, digits, it's 7LA, and that's all they know. That's what the investigators have to go off on for like a child abduction or whatever the circumstances are.

Another example is sometimes you have a combination of CCTV footage and ALPR. This is another common one I see where I posted a story a couple days ago where someone robbed a, I think it was robbed a store, and the newspaper says, a combination of surveillance or CCTV footage plus the ALPR helped lead to the arrest. Well, that's typically because what they'll see is all they'll see is a, you know, like a white pickup in the surveillance footage. They won't see anything else, just white pickup, drove off. What'll happen is the detective will go to the local, you know, the find the the camera that was flock camera that was the closest to that location, and then look for white pickup truck for like that three-minute window in which the truck could have passed the camera. And then they'll find one and then they'll go, okay, here's the license plate. And then maybe they'll find the register owner of the vehicle and go, this man’s been convicted for robbery before. This is a lead for the investigation. These are the circumstances and how these tools are used. So yeah, they're used specifically for that purpose. Unfortunately, the vast majority of that suspect information given by witnesses or or caught by a you know a doorbell camera or, you know, other CCTV is not complete enough where just the license plate itself is good enough to do the investigation.

Rafael Mangual: Yeah. So look, I, you know, I think I'm with you on on the good stuff that these cameras can do, and I think a lot of people are. And I think, you know, that's important. But I want to be fair to the critics here to the extent that I can, right? I think people who have been listening to our conversation so far, they're hearing some of the good things, some of the capabilities, right? But there are objections to this technology, and I want to talk about them for a little bit. My sense is that the objections fall into kind of two main buckets. One bucket is like the sort of legal privacy concerns, right? The Fourth Amendment violation issue, you know, these are invading your privacy, they're stalking, they're tracking you from place to place, right? And then there's the, you know, the sort of risk of tyrannical oppression, we'll call it, bucket, where, you know, if in the wrong hands or at some point in the future, right, the government's going to able to use this to oppress the populace in some terrible way. So I want to start with the the sort of legal privacy bucket. One of the biggest misconceptions, I think, with ALPR technology is that people think that it sort of follows you along while you're on the road, right? That you just put it in the license plate and it can tell you exactly where that car is, and then you can just see it traveling and finding. Now, I've been on a bunch of ride-alongs before where ALPRs have been in use. That's not really how they work, to my understanding. My sense is that, you know, yes, you get a hit from a license plate that's on a hot list, and that'll tell you which camera it passed and when. And that can narrow the search area and really increase the chances that you find it, but it does not tell you where the car is at any given point. It doesn't tell you where the trip originated, where it ended. It's not a i is it a guarantee that you find the car, or is it just a helpful tool to sort of point you in the right direction?

Rahul Sidhu: Yeah, so let's let's address those. it's not a guarantee, of course not. I first of all, people don't understand how the the cameras operate. So let's just break that down. Again, it's a a camera at like let's say an intersection that is taking a photo of a vehicle when it passes by from one particular vantage point. If you looked at every intersection, you know, each in order to get complete coverage of an intersection, you need four cameras. One that's facing in every direction. Most plates, or a lot of states don't have front-facing plates like California does, but Florida does not.

So you're going to need one in each direction to get all the license plates. And then if you think about just like how prevalent the you know, you flock cameras are in the United States. I did an now an analysis on this about a week ago. It's like 0.3 percent of intersections in America. The idea that you know you're going to be able that these are tracking everywhere you go is is certainly not true. I mean, in most communities, you they'll place the cameras specifically at the entrances and exits, as many of the entrances and exits to the community.

Rafael Mangual: Yeah.

Rahul Sidhu: as possible because the entire goal of them is, you know, if someone were to drive into that community that is like a wanted person is driving a stolen car, et cetera, then they're the police are going to have an immediate ability to to go jump on that car before it can go out there and commit the crimes. And similarly, if someone is to, you know, kidnap a child and they're exiting, I mean the the cops want that information as well. They're not on every single intersection. I don't know of a single, you know, city in America that has them everywhere where you can just be able to see where the vehicles go. And also they're taking a photo of you passing an intersection. They're not watching you pull into a driveway, you know, to know where you live or where you're, you know, shopping or whatever the circumstance is. Granted, there are some stores that have, you know, retail stores that will buy Flock cameras as well because they've become the most effective tool for shoplifting. So that the moment that a a vehicle that is tagged for organized shoplifting or is a stolen car, which is often how the organized retail theft people will operate, the moment it pulls into a parking lot, the police are called. That's a really effective way to stop, you know, shoplifters. But that's how these cameras work. That's like, you know, number one. Number two, these cameras are for vehicles. That's what they do. They take the photo of the vehicle. They don't know who is inside the vehicle. They don't know how many people are inside the vehicle. There are multiple people who could be driving that vehicle at any given moment. It's, again, that's a a big distinction from something like let's say a cell phone that is always, you know, with basically the same person at any given period of time that is pinging that device second after second and watching that person walk around and do everything they want. So that's my first answer. I mean we can dive into some of the other privacy elements and how these things are protected or the Fourth Amendment law around it.

Rafael Mangual: Yeah. Well, I mean, I want to stick with that as it relates to the Fourth Amendment law, because as you know probably better than anyone else, you know, this technology has been challenged in court before on multiple prior occasions, right?

Rahul Sidhu: Dozens of times, yeah.

Rafael Mangual: And what's striking to me is that, you know, and my colleague Charles Lehman and I have a paper coming out about this soon. But what's striking to me about the legal landscape is that there's basically, not basically, there is unanimity across the appellate sphere, meaning that every appellate court that has so far considered the issue of the legality of ALPR devices has found that those devices are perfectly legal. That is true of state appellate courts, that is true of federal appellate courts. That's pretty rare. And when I look through those decisions, the pattern that I see is that the these courts are consistently finding that there is a material difference between what ALPRs can do in the way of tracking somebody from place to place, versus say, what cell site location information or geographic historical geographic location data can do or a GPS tracker can do. Like how big of a difference are we talking about here?

Rahul Sidhu: There's a fact-based response and an opinion response. My opinion is it's pretty astronomical. But like you said, there are dozens of cases. Every every one of these cases that's been heard, that has been the outcome is that the court themes this is not, you know, triggering Carpenter's mosaic theory or, you know, any of these other things. So that is the current truth on the ground. When people say this is unconstitutional, this is illegal, they're ignoring the fact that every single court that has opined on this has come to the opposite conclusion, regardless of what they have learned on the internet from you know five minutes of chatGPT. So that's number one. Number two, in terms of like, okay, but what could happen, you know, in the future, like that's a different question and a very reasonable question to ask. I think that, like, in this country, technology is constantly outpacing regulation. This is the case outside of public safety. There are laws that were been built decades ago around medical data or financial data or public safety data. That considered at present, at the time, that might not be true in 10 to 20 years. So I think it's actually really reasonable to go, let's actually go back to the drawing board, let's talk to our legislators, let's figure out if there's a different way that we want to be able to police this data, use this data, and set new rules that might dictate what can or can't be done presently or in the future. That's super reasonable. I think that every single state in the country should understand this problem deeply, understand the balance of wanting the the law enforcement agencies to to be able to go out there enforce laws and ensuring that it's taking into consideration the rights that every single person has and the concerns around privacy that every person has because that's very important as well. And then come up with new laws, new policies, new regulations that that exist for that. In the future, then the other point to your question is, well what happens if you have, let's say, a camera on every single intersection? And let's say these devices are doing more than they're capable of doing today, and a court then hears it then. I would actually see say that if that is the case in the future, then we've actually failed legislatively. The court shouldn't have to react by then. I would rather see every state understand this technology and pass laws and kind of funnel it in in that context and go, cool, this is the way that the state of Delaware, the state of California, the state of Texas, the state of Florida is going to allow ALPR usage for law enforcement. Here are the guardrails, here's what they can and can't do.

So yeah, that's the truth. This is the current, existing, the present does not show that these cameras are unconstitutional. But I'm very big on ensuring that everyone sits there and thinks about what could happen in ten years when you apply AI to even more data than exists today. Or should we put guardrails on how many cameras can go up in a community or any of these things around the technology? And let's just set that framework so that the police can do their jobs and we don't have to have this discussion, you know, every single time a cop wants to catch a criminal when they do something wrong.

Rafael Mangual: Yeah, I want to talk about about guardrails more in in a bit, but I want to, you know, sort of stick with this like what could happen idea because I think a lot of the concerns that I have seen, you know, rest on these kind of fantastical hypotheticals of you know some dystopian future. And there's there's something that really bothers me about that. Maybe one thing that people don't know about you is that is that you were a police officer, and you know, this technology is primarily in use by police departments around the country, many of whom, by the way, have spoken out in support of using this technology, saying that the responsible use of this technology is absolutely a key law enforcement tool. And part of what bothers me about the public backlash is that it's not taking these institutions' opinions seriously or into account at all. I mean you have, you know, both Democrats and Republicans proposing all-out bans and, you know, justifying those proposals on these, you know, potential hypothetical situations. And undergirding that is basically a profound expression of distrust of law enforcement as an institution. The idea here is that these police departments cannot be trusted with this capability because they will become agents of a tyrannical state at some point in the future. I think a lot of cops that I've talked to find that offensive. And I'm wondering just just from the perspective of someone who's actually, you know, worn the uniform, what do you want to tell those those lawmakers about the people that I take them to be kind of demeaning when they when they make these claims?

Rahul Sidhu: You know, I this might sound like a hot take coming from a cop, but I actually I don't find it particularly offensive. I think it's, I think there are plenty of people out there that have a reason to be somewhat distrustful of their institutions and also law enforcement. Do I agree with them? No. But do I empathize with their concerns or in some cases their distrust? Yes. You know, and and I would say it that's kind of broad to the entire political spectrum. You could talk to somebody on, you know, I'd say left of center, who feel you know have had negative interaction to the police where they feel like you know they dealt with a racist cop. That's a pretty common one. I will tell you that not only do I empathize for that, like I've experienced it myself. People don't know this story, but when I was a paramedic before I was a police officer, I had a particularly bad run in with actually, the chief of police at the time in Pittsburgh, Pennsylvania for Pittsburgh School Police when I was driving an ambulance and this man like basically was off duty and road raged and pulled me over when I was in an ambulance, coming back from the children's hospital, and like started yelling at me because he said I'd cut him off in traffic when I was on my way to a call, which is an insane thing to do. And when you know, he had to he got punished for that and I had to testify against him in court, and he told the court that the reason he pulled me over is because he thought I was driving a Taliban war wagon.

Rafael Mangual: My goodness. Wow.

Rahul Sidhu: And he was protecting the city from a cloned ambulance car bomb. And that's an insane thing to do. That guy got fired, obviously. And that was me testifying against what I think is like a racist police chief and a person who shouldn't be a cop. And there are cops out there who shouldn't be cops. That's reasonable. Now the vast majority of them are great. Like, don't get me wrong, but of course, just like any profession, it has its bad apples. On the center, I'd say right of center side, you have a lot of people who were frustrated by either their local community's response or the state's response or the federal response to COVID back in 2020. And I understand their frustrations. There were things that ended up, you know, being not true about COVID and they were particularly frustrated with how the government reacted. I don't, I can't look at those people and say, like, always trust every single thing that you hear.

I can tell you that I was a reserve police officer during COVID, and our community policed it, you know, things differently than other communities. That's the local response. We chose a pragmatic approach to how things, you know, should happen. And it worked very well for our community. And our community is particularly happy about it. But I also saw communities in the country where, you know, like some guy would get a ticket for surfing, you know, which is like in my opinion, pretty unreasonable. So when people have like, you know, concerns about that, I would understand their concerns. Where I think that my opinion differs is despite the fact that that is like a reasonable take, from a practical standpoint, their concerns around ALPR being used for that context, I don't agree with. And it's for a variety of reasons. Like number one, if you just think about the context, let's just stick with COVID for a second, of enforcing a lockdown. I don't know of any community in America where the lockdown was you simply cannot drive on the road. You just cannot, it doesn't matter what you're doing, you cannot be on the road. I never heard of anything like that. That's the only context in which I think an ALPR could be used to enforce something because if you're driving on a road, you pass the intersection and go, cool, this vehicle is driving. Let's, you know, even then I don't know who's driving, so I can't really write a ticket, but you know, there's some something there. The context of like you go to, you know, I'm driving to a barbecue with four people who aren't socially distancing, like an ALPR is not going to help with that. Now, if you think through this like legally speaking, let's say the federal government, some president that we don't agree with, becomes the president of the United States and decides to enforce the craziest lockdown known to man and goes to a local community. Like my old police department or some other agency and says, hey, we want you to give us all of your ALPR data. Let's say we have six cameras in the in the city, very you know, not a crazy deployment, and we say, no, we're not going to do that. And they go to Flock and they say, “Give us all your data,” and we go, “No,” we're not going to do that, and it's not a lawful command. The federal government does not have a lawful reason to have it. We say, “No.” If that particular federal government goes, “We don't care, we're going to take it by force, even if it's illegal for us to do so,” we have a much bigger problem. This is like, ALPR is no longer even in the top one hundred things that we should care about. At that point, they can go to Apple and force them to give all that data. They go to the cell phone companies, the data brokers, they can go to Tesla or your Ford and say, where was this vehicle? Like that is way worse. If the concern is that the local police department is going to be used for the Gestapo in this context, they have guns, they have armored vehicles, they have all other databases, they have all these other ways if they're going to use these things unlawfully to do these things. So we're talking about removing tools from law enforcement with practical guard laws already in place for this type of behavior and being concerned that this one thing that couldn't really be practically used for the context of a lockdown might be used for this purpose. Of course I empathize with the concern. I just don't think it's rational in that context.

Rafael Mangual: Yeah, no, I'm with you. I mean one of the more frustrating memes around this whole debate is this idea, you know, this question rhetorically posed of, you know, imagine if we had these things during COVID. It's like we don't have to imagine. You know, ALPR technology existed.

Rahul Sidhu: Yeah, it existed. There were 70,000 ALPR cameras deployed nationwide in the United States, a little over 70,000 in 2020. Two out of three, so two thirds of major cities in America by 2020 already had ALPR programs. This is not a hypothetical, this already existed.

Rafael Mangual: Right. One of the things that people say in response to that is like, well, there weren't as many or you know, the cameras weren't as capable, but I've yet to really see anyone make the connection between how the capabilities change and why that makes, you know, tyrannical oppression more likely.

Rahul Sidhu: You could you could make the argument that let's say Flock has a hundred and thirty thousand now and there were seventy thousand before there's almost twice as many ALPR cameras. Does that make it does that all of a sudden go from zero to one? Is it like a binary now it's a problem, but it wasn't a problem then? You could make an argument that AI generally has gotten better, the plate reads are more accurate, or the vehicle descriptions are more accurate. Is that a binary step where all of a sudden it couldn't be used for this hypothetical oppression, but now it is? It's subjective, I guess, but again, even if you had today's capabilities, that's kind of my point, and you applied it at the scale that we're at today with the capabilities we have today, how? How is it going to be utilized in the context that folks are concerned about? Like nobody can really answer that question outside of, you know, yeah, the vehicle was on the road at this date and time.

Rafael Mangual: Yeah. I want to talk I want to shift a little bit into safeguards. You know, a lot of people who are upset about these things or who are expressing concerns or maybe on the fence. One of the things that they often point to are these examples of police officers abusing their access to these ALPR systems, including Flock’s, where you know, you'll see these stories of you know, an officer who who looked their ex up, you know, a hundred and sixty times in a week or, you know, who, you know, tracks down their daughter's boyfriend's car and, you know. Now I see those cases and my first thought is, and I've been in this space for a long time. I mean, I've been writing about policing and public safety for well over a decade now, and I see those arguments, I think back to 2016 when everyone was talking about this Associated Press expose that looked at police departments around the country that had cases of officers abusing their access to DMV databases and background check systems, you know, to run down their daughter's boyfriend or to look up their exes. And it was the same arguments being made back then. We seemed to to resist it as a country a a lot better, but I remember thinking like, well yeah, you know this this is going to happen, you know, but but these are statistically isolated incidents that no more justify removing the tools than statistically isolated incidents of excessive force justify, you say, you know, disarming police.

Rahul Sidhu: Yeah, look, you're talking about the statistics, so I'll I'll start with that. If you have, we have like a little over 140,000 monthly users. and if I just randomly said—I'm not saying this is true—but let's just say a hundred of them have been caught misusing the platform, it's a pretty small percentage. But I don't want to, I don't want to, I don't even like saying that because all of them like any amount is bad. all of them who have been caught, and we can prove that they actually did this, like my opinion is they should go to jail. That's a bad thing. The reason we hear about them is because Flock’s auditing was able to catch them doing it.

Rafael Mangual: Well, this is what I want to talk about, right? Because like, you know, people talk about safeguards, you know, kind of in a hand waving way, but I mean it sounds to me like when you…

Rahul Sidhu: This is the safeguard. Yeah.

Rafael Mangual: Yeah. So like describe it to me, right? So how does, what is it about your system that makes it possible to catch misusers?

Rahul Sidhu: Yeah, so look, there's there's a variety of things. and and to be fair, some of these things we've rolled out over time. It's not like these have all been in existence, you know, from the beginning. And I think it's a fair criticism to wish that they were all in existence from the beginning. We are currently I would say like out of all the police databases out there, this is my opinion, we're the most sophisticated when it comes to all of the guardrails that are in place, not just for all the ALPR companies, but just generally with police databases, because it's going to understand what misuse looks like. Number one. I'm talking about today, presently. Case numbers, for example. We announced that all the case numbers are required. There needs to be a lawful law enforcement reason that's documented. I want officers to do a better job actually explaining those in long form because I think that's like one thing that's like difficult. But I, you know, would agree with criticism there. There is, if misuse is detected, let's say an officer is constantly, you know, looking for a plate and coming up with different case numbers, you know, stalking their ex. That misuse does get detected.

Rafael Mangual: So when you say case numbers, just walk the audience through that, right? So an officer who's logging into the system has to create a unique case number for every search.

Rahul Sidhu: No, they don't create a case number, they reference a case number. So…

Rafael Mangual: Got it. So that has to be in relation to an ongoing criminal case that has a number within their system.

Rahul Sidhu: Correct. So it's like, I'm investigating for this case that is in my investigations management software. The agency has established this case number for that case. So I'm going to put that case number into the search. So if anyone audits a search, they know what case this particular search is in reference to. That's one of the most effective ways to to determine if a, you know, a search, to audit something and go, what was the context of this particular search?

Rafael Mangual: And is there a way to know which specific officer is actually doing the search? So beyond the case number, is there like a personal ID?

Rahul Sidhu: Yes, yeah, there's an employee ID. You know, what you're describing essentially, I'm simplifying it, but yeah, you know who did what search, at what date and time, the case number, the law enforcement reason. And there's also, you know, we announced this recently, but proactive lockout. You know, I'm going to I'm going to give kudos to —this is going to be kind of a hot take —but we have critics out there, that, some of whom are journalists, you know, and content creators, et cetera, cybersecurity researchers. I'll even name them by name. Guys like Ben Jordan. I might be getting the name wrong. I see him every once in a while. He's an Atlanta journalist. Brendan, Brendan O'Keefe. These are guys who dove into audit logs and found discrepancies and I would say like exceeded what the law enforcement agency themselves were doing, to be able to determine that misuse was occurring. And I, you know, I actually applaud them for doing that because it shined a light on the fact that some agencies might not be doing a good enough job actually doing that auditing. We actually ended up making audit assistance mandatory for every agency.

Rafael Mangual: Audit assistance meaning what?

Rahul Sidhu: Rafael Mangual: Got it, got it.

Rahul Sidhu: Several decades. Most police databases, like they have some auditing capability, but it's nowhere close to what Flock is able to do today. So like a very common problem across the United States is a cop will get a, you know, a call from a buddy, and that buddy will be like, “Hey, can you run this plate for me?” And the cop, “yeah, sure, buddy.” And then run the plate. “Hey, it comes back to this.” And that's like the DMV database. And it turns out that guy was calling his cop buddy because someone, you know, hit on his wife at a Trader Joe's parking lot and now he has their name and address. That's terrible, right? And then you have circumstances where cops will run the criminal history of their daughter's new boyfriend, or they'll find an attractive woman driving down the street and they'll run her license plate in the records management system, get her name and then go follow her on Instagram. And these are terrible circumstances. The auditing for those systems is not nearly the same. So the vast majority of times it happens that cop doesn't get caught. And so we announced recently like, look, when we put out this feature, we had this discussion where we said, we're going to basically force daylight on all of this. And it's going to catch a bunch of cops who otherwise don't even like think that they're doing something wrong in some of these cases. And they you know, they're running plates or whatever the circumstances are. It's going to catch them, it's going to shine daylight on them, and it's going to get them punished. And we were hoping, of course, that you know, we're going to go out there and do the right thing. We were hoping that there'd be some credit for this, like, okay, cool, we're going to move and then everyone would follow us in the police industry. Let's just rip the band aid off and move towards the more accountability, more transparency approach on police databases. So far, I don't think anyone's really followed us, unfortunately. And the problem is that basically every news story is about Flock, it makes it seem like Flock is bad because it was being misused, versus the fact that the only reason you hear about all this misuse here and not for anything else is because we're the only ones shining daylight on it. And it's a really painful way of trying to move the, you know, the culture and the the accountability and transparency in the direction that I think everyone can agree is ethical. But even some of our critics like kind of shining a light on these things, like that's actually a useful movement towards making that the case. Beause I think now every agency is going, wow, we really have to audit and police this stuff. And that's a useful thing. We should want that. And I'm saying that as a cop, like we don't want cops to misuse any of these things.

Rafael Mangual: Yeah. No, I think that's important. I think, and you know, I don't unfortunately expect that you'll get a ton of of credit for for taking those steps in the public. I mean, one of the trends that I've, you know, talked about for a long time in policing, right? I mean, use of force is a great example of this, right? In 2020 you saw this sort of explosion of anti-police sentiment over, you know, alleged cases of excessive force, and, you know, in my attempts to kind of contextualize those things and the broader sphere of police data, I would always point out that look, man, you go back to the 1970s and look at the use of force numbers back then. I mean, the NYPD shot, I think, more than 200 people in 1971. They killed almost a hundred. Now officers discharge their firearm maybe two dozen times a year in a city of eight million people, right? Which is much bigger than the population was in the early 1970s. And yet you don't see any sort of credit for for that trend reflected in the public sentiment, especially within the the sort of protest-driven left. So yeah, I mean, all that to say don't hold your breath for a ton of of credit for for doing you know the right thing necessarily, but I I'm glad you guys are doing it. So…

Rahul Sidhu: Yeah, look, the vast majority of cops, they're just people trying to do the job right. And they're out there following the rules. And it's like it's a very, very difficult job to do because most of them are underpaid, in my opinion. Most of them are under resourced in terms of their tools. They're, you know, they don't get the training that they would otherwise want and ask for, and a lot of them work for communities that don't even want them to exist in the first place, which makes it really difficult. And they'll go out there and they'll do this difficult job and then we'll ask them to be better and do more and and they're not, you know, catching enough guys, bad guys and they're not, you know, doing enough. And it's a mentally very draining job. So I don't want to, you know, shine a light on, make it seem like this is the case for all of them. It's a very small percentage of people who are misusing these databases, a small percentage of people are doing the wrong things. And every cop that I know, all the good cops I know, like they don't want those cops around. They don't want them to bring every other cop down because you can imagine, you have some agencies where they'll find out, you know, one of our cops misused the database and great man, you just ruined it for everybody. That sucks. That person gets yelled at internally and fired and and hopefully goes to jail and sets an example for what not to do.

Rafael Mangual: Right. Right. Yeah, no, I think those are all important points to be made. So, I mean, one of the reasons I wanted to talk about safeguards is because I see them as a way to kind of mitigate the risks that people are concerned with when it comes to ALPR technology and you know surveillance technology more broadly, right? You know, not just ALPRs, but CCTV feeds, drones, all of that sort of thing. I think a lot of the backlash is rooted in this idea that there is risk. And safeguards can help mitigate those risks, right? And the reason I think that's important is because every policy question, and ultimately I do think this is a policy question, it really comes down to do the costs outweigh the benefits or vice versa? And you know, we've talked about like the potential risk associated with the use of this technology or the technology falling into the wrong hands or being misused or abused. But I want to end our conversation with just a reiteration of what the benefits of this technology are.

I was reading, you know, when I first kind of, not really first, because I've been talking about the importance of technology and policing for a while, but when I first sort of got engaged in the most recent you know controversy around your company, one of the first cases that I thought of was a recent case of a woman whose car was stolen with her two-year-old in the backseat, her two-year-old daughter, I believe. I think it was in Atlanta. The police were able to I mean, I heard the frantic 911 call on the news and it was just, you know, it's bone chilling, right? Because I'm a parent. I have two kids. You know, you think about, my God, like, you know, what would that be like? And I think they caught the guy within fifteen minutes and recovered the girl safely in a parking lot using this technology to sort of triangulate the area where the car might be found. What are some of your sort of favorite stories or you know, use case examples that that you think sort of justify the maintenance of this technology, even given the potential downside risk?

Rahul Sidhu: You know, that case in particular, the one you're referring to, was from a few weeks ago in Atlanta. That one frustrated me when it was posted online because everyone kind of resorted to victim blaming this lady. I think she was like delivering DoorDash or something. And so she, you know, which like, just if you're just closing your eyes and picturing that like this poor woman is having to like take care of her child while making ends meet on a gig economy app, like DoorDash, like that kind of sucks. And I feel bad for her there. But she's, you know, out there doing what we expect for her to do in society, which is like work really hard, try to take care of her child at the same time. And as you can imagine, DoorDash, sometimes you have to get out of the car and go deliver something or pick up food. And that was the context in which someone immediately ran and and stole her car with a child in the backseat. And so many people responded, Well, this is just like a bad mom. Why would she leave her kid? And like and that that was like that really bothered me. But look, in that context, yeah, they rescued the kid. I saw the body cam footage. It was incredible. They they caught this guy who I don't you know, he had priors. I don't think he should have been like free to have committed that crime in the first place just given the circumstances.

But there are a lot of stories like that. A lot of stories. I mean, I'm not talking about like a few. I'm talking about hundreds or thousands, over the course of the last several years where ALPR was used specifically for a kidnapping. And a lot of people go, Yeah, well, how many of those kidnappings were from family members or strangers. Well, let's just think about that for a second. A lot of them were from strangers more than people think. But even if they are from family members, that doesn't inherently make them okay. And to say like, we know who they are, so you don't need the license, that's not how it works. When a dad decides, okay, I just lost custody of these kids because the court deemed that I am deranged. So I'm going to steal the kids and try to drive down to Mexico before I can get caught. Just because you know the identity of that person does not mean you're going to find that vehicle. If an automated light slate reader in like Escondido, California, or Chula Vista, California, catches this guy on the way down and it's an Amber Alert and we stop him, that's a good thing that could have potentially otherwise not happened before he crossed that border. So that's like number one. I think there are a lot of cases, like if I broke down some of the best examples, there's a difference between circumstances like that, which I think is like a really important circumstance where you're actually rescuing and potentially saving the life of a child or an elderly missing. Like I think those are my favorite ones because that is a like potential life saved. Then you break it down to a different category, which is a crime solved or someone who actually, you know, went to went to jail because they were caught after they did something heinous, whether it was a murder or a rape or whatever the circumstances, something bad to a child, they got caught because ALPR caught them or ALPR was used as evidence to convict them. And in cases where they might not have otherwise gotten caught because the ALPR was not there, or they might not have otherwise had enough evidence to tie that person to the crime because the ALPR did not exist. I'd say that's like my second tier of like that's particularly useful. The other aspect of this where I think it's useful, I have like a pinned tweet on my x profile, because this one's really important to me.

Yeah, is just like let's just break this down from a common-sense standpoint. Imagine two cities, a tale of two cities. You have city A and City B. In City A, they have automated license plate readers on like, let's say, every entrance and exit to the city. That's it. Not everywhere else in the city for the residents, but just the entrances and exits. And they have drone stations placed in the city with the, you know, with like Flock rapid response drones. and every single time someone drives into that community with a stolen car or a swapped, stolen plate, or you know, they're wanted, or whatever the circumstances, the moment they drive into that community and that city in City A, the license plate reader goes off, notifies police, the drone takes off and goes and finds that vehicle. And every single vehicle that drives in to go commit crimes in that city is caught before they can actually do so, let's say in a matter of 60 to 90 seconds. And then you have city B that does not have any of that technology. And you're a criminal and you're going to go into which city to commit the crimes? City A or City B? City A, your likelihood of getting caught is astronomically high. And in City B, your likelihood of getting caught is much, much lower. And when we think about what deters crime, there are actually a variety of valid approaches. The one I'm referring to is if the likelihood of getting caught is too high, then you're probably unlikely to commit the crime. There are people who have a valid argument, I would say. I would say it only takes a piece, a piece of the equation into play, but they have a valid argument where they go, yeah, but in my community, we're constantly catching very serious offenders and we're like letting them out too early. That's yeah, I have an individual political opinion on that. Like that's an issue by itself, and that's another portion of how to deter crime is yes, it's great that we're catching people, but after they're caught, if there's no consequence or the consequence is very little, then that's also an issue. My opinion is that we need to solve both those things. But if you don't catch them, you cannot have consequences in the first place. So I think that first thing is actually more important in some cases than the second thing.

I'll end with this. You know, I remember this story when I was a like working on patrol as a police officer, I called my jeans guy story, where this guy he was, it was a shoplifting call at a retail store where he stole like a dozen jeans. I don't remember, it was like some astronomical amount of jeans that he just grabbed and then he just walked out of the store. Doesn't matter, just walked right out of the store and started walking towards a bus stop. And my partner and I, we see him, we sit him down, we're like, “Hey man, we know it's you. You have, you're the only guy carrying, you know, twelve plus jeans with you.” And he goes, “Yeah, yeah, you caught me.” And we're like, “okay,” so I take his jeans away. I'm searching him, I'm getting ready to at that point, which is really frustrating, it was just a citation at the time where we would write him a shoplifting ticket and I'm searching his his clothing and I find he has two different shoplifting tickets similar for different jurisdictions that are near the area. And I go, “What are you doing? Why do you do this?” And goes, “I steal the jeans and then I go to the flea market on every Saturday or whatever. And then I give them to a guy who gives me a bunch of money for them and I get paid all this money.” And the county was at the at the time not processing this guy in terms of like a warrant because it didn't exceed the threshold for us to process him. So he would accumulate all these tickets. They wouldn't care about the cumulative value. So if we were to try and arrest him on a warrant for not showing up for court for any of these, the county would just turn them away. So we'd have to write him another ticket. That's a reasonable thing to be frustrated about. But if I never caught the guy in the first place, I still can't solve that problem.

 So I'd say that's like my tiers for all the ways that I look at how this technology is useful. And then if I just think to myself, like if you have this city where it's completely deterring all this crime, that's kind of the utopia we want, right? Like if there are guardrails in place where you're not victimizing innocent people, you're ensuring that every single time a police officer gets a hit, for example, on a vehicle, he is verifying that the hit is accurate and he's verifying that the database has the appropriate vehicle in there. And then he, you know, does that traffic stop, you're going to completely cut down your possibility of having incorrect traffic stops and you're routinely just never going to interact with members of the public that are not committing a crime or stealing a child. So I think if you think about that in its context, it's a very positive circumstance as long as the downsides are considered and there's appropriate safeguards.

Rafael Mangual: Yeah, no, I have to say I agree. I mean, you know, when I think about, you know, that critique that you just highlighted, this idea that, well, on the back end, we don't do enough to, you know, punish these guys. I mean, that is a critique that I'm very sympathetic to. I wrote a whole book about why we needed to incarcerate more people and punish people more harshly and and pursue prosecutions more aggressively. But the reality is is that you know, clearances have been a problem in America for a really long time. I mean, historically, clearance rates for violent index felonies have been below 50 percent. Historically, clearance rates for property index felonies have been below 20 percent, meaning the vast majority of the time that people commit crimes that are reported, right? Save the fact that most crimes are not even reported to police, but the vast majority of the crimes that are reported don't result in any punishment. That is a problem. Raising the likelihood of catching people is a possible solution to that problem. The more people you catch, even if the incapacitation is only temporary, right? Even if they only spend a weekend in jail or a few months…

Rahul Sidhu: Yeah, there's still a downside to it. They don't want to be caught. Yeah.

Rafael Mangual: Exactly. Right. You know, that still matters. So look, I could talk to you for another two hours, you know, if I could, but I appreciate you being generous with your time. I appreciate you coming on the show, talking about these issues, which I think are not going to go away anytime soon, but I think you know that the that that you and your competitors are doing is important and I hope that society eventually comes around to that. I think they will. So again, thank you so much for joining us today.

Rahul Sidhu: Thanks for having me on.

Rafael Mangual: Awesome. And for you all watching, this has been another episode of the City Journal Podcast. Please do not forget to like, comment, subscribe, do all the things. Let us know what you thought of the episode if you have any questions, follow ups, that sort of thing. We look forward to seeing you all again soon.