A PSA on ALPRs
2022-09-21 · Guest: Jonathan Hofer (Independent Institute) · 52:45
Automated License Plate Readers and Surveillance Privacy
Bob Zadek interviews Jonathan Hofer of the Independent Institute about the pervasive use of Automated License Plate Readers (ALPRs) by law enforcement. They discuss the technology’s mechanics, the lack of privacy protections, and a harrowing personal account of a “high-risk felony stop” caused by outdated data.
Topics: ALPR, surveillance, privacy, Fourth Amendment, law enforcement, mosaic theory, Independent Institute
Speakers: Bob Zadek, Jonathan Hofer
Introduction to ALPR Surveillance [00:00]
Bob Zadek: Hello everyone, this is Bob Zadek, welcoming you to the country’s longest-running libertarian broadcast, available 8:00 AM Pacific Sunday mornings on select AM stations and streamed nationally at that time through 860 AM The Answer. A decade of prior shows are on the Bob Zadek Show podcast and BobZadek.com has related source material, book lists, and other podcasts of interest. We offer in-depth and focused content on social, political, and economic issues, always with the ideal guest, accessible and entertaining. In short, ideas, not attitude.
Today’s show embodies that standard. To paraphrase a scary old public service announcement I used to hear on television—remember television?—and even on radio, it was, quote, “It’s 10:00 PM. Do you know where your children are?” To modernize that public service quote, the 21st-century local governments, particularly their police departments, could boast right now, whatever time it is, “We know where you are.” Unfortunately, that’s probably the case.
Today’s guest, Jonathan Hofer, is a research associate at the Independent Institute based in Oakland, California. He has studied privacy law, local law governance, and the impact of emerging technologies on civil liberties—and what an impact that is. Jonathan will introduce us today to a tool which adds teeth to that theoretical quote by police departments: “Whatever time it is, we know where you are,” and perhaps I should add, “and where you’ve been.” Jonathan, welcome to the show today.
Jonathan Hofer: Well, thank you so much. Thank you for having me, Bob.
How Automated License Plate Readers Work [01:34]
Bob Zadek: So Jonathan, is it an exaggeration for me to frighten our audience into believing that local governments around the country—and I should add, California once again is blazing the way, but this is true nationwide, this is not solely a California problem, it is absolutely a national problem—is it an exaggeration to say that local governments generally always know where we are? And if that’s true, how do they learn that?
Jonathan Hofer: Yeah, I think we have a lot of reason to be concerned. You know, it’s kind of interesting post-Edward Snowden leaks, I think a lot of attention in the context of government mass surveillance, we usually think of things like the CIA and NSA. But in reality, in terms of numbers, the most prolific tool of mass surveillance today are automated license plate readers. And even though they’ve only been around for a few years now, they’re quickly exploding onto the scene in the United States. And if your town doesn’t have one, it’s soon coming to a place near you.
Bob Zadek: Automated license plate readers. Well, seems a little bit bland, a little bit uninteresting. So somebody reads license plates. People read all kinds of stuff. What’s wrong with reading a license plate, or is there more to the story than that?
Jonathan Hofer: Yeah, it’s really kind of an interesting issue. So automated license plate readers, or usually what are called ALPRs, are specially designed high-speed cameras that have the ability to read the alphanumeric characters on a license plate. And they use it—they are able to read the numbers and letters using a special software called optical character recognition. And I kind of use the comparison: it’s a similar technology that allows you to keyword search a PDF that you might have scanned, or if you’ve ever gone on Google Books, it lets you search printed books. Same kind of technology.
They were originally developed around the 1970s in the United Kingdom, and then they started popping up internationally. And for many decades, they were not advanced, they were not practical. To the extent that they were used, they were only used for things like regular traffic monitoring or sometimes they would be used for electronic toll collection or parking enforcement. But around the early 2000s and especially into the 2010s, they quickly exploded with law enforcement in the US when police started surmising that this would be a really good crime-fighting tool.
And so how they generally work is that they’re stationed at either busy intersections or over a busy roadway, or they could be mounted on a police cruiser or other city infrastructure. And when a car passes by the camera or the camera passes by a car, the license plate is detected from the image and a timestamp is applied to that scan, and then there’s also basically a GPS coordinate; they log your latitude and longitude. And this poses a number of civil liberties concerns.
Data Aggregation and the Mosaic Theory [03:55]
Bob Zadek: Well, let’s back up, Jonathan, if I may. Sorry to interrupt. But to follow it sort of sequentially: we first have a camera, to put it simply, which records the information you just described. Now, so far, so good. We just have data, rather benign. But you started to say how it gets used. But there’s a step before that where the data is organized or accessible or it’s collected from many locations so we can not only tell where you were—you being the car, we don’t know who’s in it, but we do know the car has been in certain places, which gives us a clue but not proof of the person. So we know where the car has been. But tell us about the story a series of images tells about the car and presumably then about the driver if we know who the driver is. So tell us that story in gross. What is the data that now is being collected and who keeps the data? Is the data like available at a website on Google? So tell us about who’s holding the data and what the data might reveal about the car. Not about the driver, it’s not about humans, not yet at least. So tell us that story.
Jonathan Hofer: Yeah, I’ll actually say it could be the bleak reality that it is the driver and humans because usually these cameras try and take at least six pictures of an individual car, and that’s per manufacturer’s recommendation just to make sure the license plate is legible. So that can include the occupants of the car being photographed.
So when we’re talking about this license plate data, we’re not just simply talking about the storage of the literal license plate number. What we’re really talking about is a roadmap of where you’ve been, where you’ve traveled. And when you have a series of these cameras, you can learn a lot about people’s lives. And usually when a camera scans a license plate, that is uploaded to a centralized database. Now it varies by jurisdiction to jurisdiction who has access and who retains this data. There’s usually three kind of types of data storage arrangements. It could be an individual department stores their own data—for example, maybe your local police department or the California Highway Patrol. It could be a private third party. So for example, in Northern California, there’s this company called Vigilant Solutions. They get their own records and then they make it available to state and federal agencies.
Bob Zadek: For a fee.
Jonathan Hofer: Yeah, for a fee, of course. And then there’s also intelligence sharing bodies. So usually these are organized under the banner of the Department of Homeland Security. They kind of pop up post-9/11 as counter-terrorism centers. They’re usually what are referred to as fusion centers. And what they do is they aggregate data from multiple sources and then they also make that available to other agencies, other states, both federal and local.
So the big concern in terms of just general civil liberties is that when I have your travel records—and I should remind the audience that, you know, this isn’t just like we’re randomly spotting license plates. In the case of Los Angeles, for example, they’re retaining this data for upwards of five years. And if I have your travel data for five years, I really have an intimate picture of who you are. I know where you work, I know where you live, I know what bar you frequent. I could even know where your place of worship is. I’ve even made the point to other people, you know, I’m sure the federal government would like a national firearms registry, but we can infer if you have a firearm if you’re traveling to a gun show or to the gun store or firing range, which are usually remote. That’s good enough for them. And in fact, we have instances of the state police in Virginia doing just that. So it really creates a massive Fourth Amendment concern once you’re collecting this data, having a large historical data set that you can analyze for individual people. And especially, this isn’t just for people who are accused or suspected of criminal wrongdoing; this is anyone who’s just driving on the street.
Privacy Rights and Public Spaces [07:59]
Bob Zadek: One thing you mentioned a few seconds ago, I perked up—I mean, I perked up over all of it, I haven’t been napping, Jonathan—but one, you said in Northern California, I believe you pointed out that the data collection was done by a private party. So first of all, I imagine since I am allowed—I’m a private party—I’m allowed to have as my hobby standing on the freeway taking pictures of license plates. That’s not against the law. So I presume that there’s no law that regulates the collection by as many private parties who want to collect the data if they have the right to put their cameras on poles or whatever they need. So in the first instance—we’ll get to law enforcement in a moment—but in the first instance, since we know that data is today a commodity, it’s saleable, it depends upon how it’s packaged and how it’s organized, but data per se, facts, are a commodity which are routinely bought and sold, starting from mailing lists and going on from there.
So now we have a private party which gets the right by leasing or easements, whatever it needs to get, to put a bunch of cameras all over the place. And it collects the data. Now, that private party, are they or are they not themselves free to sell that data? Or are there privacy regulations that regulate you and I from buying a whole lot of cameras, putting them on poles, collecting the data, and then say, “We got some data about who’s driving on this freeway and we’ll sell it to you”? Is there any regulation or law governing that activity, or is that not—does that not implicate the right to privacy you referred to a few moments ago?
Jonathan Hofer: Yeah, that’s a great point. And it’s a really interesting legal question. So by and large, since the late ’60s, case law in the United States when regarding surveillance topics has largely focused on the issue of if you have an expectation of privacy, and that’s in the Katz decision, I believe in 1968. And one of the things that we talk about when we’re looking at if you have an expectation of privacy is, is it just like in the plain view, is it in the public square? Obviously, roads are public. And when you put those together, you don’t really have an expectation of privacy over your license plate. And in fact, the Supreme Court has explicitly addressed this topic and they’ve explicitly said that no, you don’t have an expectation of privacy for your license plate. It’s not illegal to photograph a person’s license plate. You know, you could set up your camera on the roadway. But it does start to broach into another Fourth Amendment legal territory once you start aggregating the data.
Bob Zadek: Well, let’s back up just a moment. Right to privacy, the concept of right to privacy describes the relationship between individual and government. Right to privacy doesn’t describe relationship simply between private parties. So if you and I, who are bored and entrepreneurial, decide we want to—we have developed a business model. We’re not talking about law enforcement, we are not law enforcement, we’re not deputized, we’re just two guys with time on their hands. And we put up a bunch of cameras and we start to collect the data to build an inventory which we hope to sell, because we hope somebody will want to buy pictures of license plates going down the freeway. If I stop there and we take our information and we make it available, we package it, and it’s on the internet and for $19.95 you can buy this data, something like that. Anything wrong or illegal about that, or is that—does that not implicate the right to privacy you referred to a few moments ago?
Jonathan Hofer: You should actually be totally in the clear. There’s no state or federal laws that would prevent you from doing that.
Bob Zadek: Okay, so the collection and organization of data by a private party simply as a business activity is—we’re not talking about that per se. It’s only when government invites itself to the party, if you will, and wants to take advantage of this data. Now we implicate the right to privacy that you have so accurately described a few moments ago. Okay, so now back to our story, Jonathan. We now have the data being collected and aggregating, and there are organizations, governmental and perhaps private, that will take the data from one county and then for the next county and before you know it you have the state, and then you can weave it together and through the miracle of digital magic, you then can isolate a license plate from all of these pots of data that you have collected, and now you know everything about a human being you’d ever want to know.
Jonathan Hofer: Absolutely.
Bob Zadek: Okay, so now we know what the technology is. We know what can be done. Well, okay, that’s what can be done. But something peaked your interest at the Independent Institute which caused you to write a somewhat alarmist—it was well-tempered, be sure, because you’re a calm guy—but it was, “Hey people, there’s something going on that you should be aware of.” Sort of like Harold Hill about trouble in River City, but your issue is far less benign than a pool hall. So tell us, tell us now why this gets you agitated.
Pulling Over the Wrong Person: Jonathan’s Story [18:08]
Jonathan Hofer: Yeah, so my interest in license plate readers comes from really a personal story. Back when I was an undergraduate at Cal, my brother and I were traveling back to Oakland in a rental car from a Thanksgiving break. And almost—we’re just a few miles away from home and we get the Contra Costa Sheriff’s Department flashes his lights behind us. And we’re like, “Okay, what are we doing? Like, we’re not speeding, we’re following the traffic laws.” And then it gets even weirder when he gets on his loudspeaker and says, “Exit the freeway.” Right? So this is not normally what you’d do for just a routine ticket.
We end up in what’s basically a vacant shopping center parking lot. It’s a November night, it’s basically pitch black. Some time passes, but we notice that one of the officers gets out of his car, has his gun out, and he does the thing where he says, you know, “Take the keys out of the ignition, put them on the roof, put your hands through the steering wheel, put your hands outside the window.” Some more time passes and then now we’re completely surrounded. So at that point, maybe six, seven, eight other officers are surrounding our car in probably about four or five different vehicles. They all have their guns out. They tell us to exit the vehicle separately.
And when it’s my turn to get out of the car, I hear an officer, he’s telling me, you know, “Take so many steps backwards, take so many steps to my left,” and so forth. Out of nowhere, I get tackled from behind. Just like, absolutely—you know, it’s like a linebacker or something. This wasn’t a friendly handcuffing. And in that process, one of the sheriffs just puts a gun to the back of my head as he’s just yelling at me. It was totally incomprehensible. I followed their directions very explicitly. I was fully cooperative. I wasn’t saying anything.
And after that, I was put in—they separated us out into two separate cars. About 45 minutes passes and they said, “Hey, we’re sorry, but we thought your car was stolen. One of the automated license plate readers that you passed had identified your car as stolen, and when our deputy pulled up behind you, it flagged your vehicle again.” And what had happened was that the car we were traveling in was reported stolen in San Jose like a long time ago. It was totally recovered. The police even call up the owner of the vehicle because it was a car-sharing rental. And they said, “Hey, we got your stolen car.” And the owner’s like, “What are you talking about?” And what had happened was the San Jose PD never updated the stolen car registry. And so the Contra Costa Sheriff, acting upon outdated information, turned our traffic stop into a high-risk felony stop, which is usually police procedure to handle ALPR flags like this.
So I think my situation was that this could really happen to anyone. And I kind of looked and saw if this had happened to other people. Sure enough, this has happened all over the nation. Just a year or two prior to my case, a woman in San Francisco, she had just settled with San Francisco PD over a similar stop where they pulled her out at gunpoint. They even had guns on her after they handcuffed her. And they alleged she was a car thief. But her burgundy Lexus wasn’t the gray GMC truck they were looking for. They didn’t even check the make and model. But there’s a real serious risk when relying on this inaccurate technology. And sometimes it’s just a failure on the camera’s parts, but sometimes it’s also just a failure of accurately maintaining the data. But you know, this is commonly pooled data. So many departments are contributing to these databases and one error on the part of a single party really magnifies.
The Effectiveness of ALPRs [22:49]
Bob Zadek: What are the—so there’s a risk of error. Now, of course, one—if an individual or a listener to our discussion was less sensitive to violations of personal freedom, violations of a right to privacy perhaps if that’s the implication, isn’t a response, “Well, of course, in every aspect of life, sometimes there are mistakes that are made. In fact, police often—well, sometimes, maybe hope not often—they arrest the wrong person. And they discover the error and they do the best they can to apologize and whatever they do when they arrest the wrong person. And that’s not an argument that police shouldn’t arrest people. That’s an argument people should be more careful, or that people who are victimized perhaps should be compensated. Or not, because if it’s compensated, then it means all other taxpayers who didn’t—had nothing wrong with the wrongful arrest have to pay. Well, that’s part of society, that can be absorbed by society.”
So are you making the argument that there is something inherently evil or at least anti-civil liberties about the technology and the process, or are you making an argument that, okay, there is lots of good that comes from it—recovering a stolen car is good, finding somebody who’s on the lam is good, that’s good for all of us, we are safer—and therefore the technology does a lot of good, so let’s focus on the bad and mitigate it? So what exactly is the headline of the message you wish the audience to have when they come into contact with or read about or are voting for for funds for automatic license plate readers?
Jonathan Hofer: Yeah, great point. Broadly, I would say that I am optimistic that given sufficient safeguards, ALPRs could be used to help law enforcement address crime. However, as they are used in California especially and how they are used nationally, I do not believe they justify their current use. I believe without safeguards, they are an inherent civil liberties risk. And I believe the benefit to law enforcement is minimal to non-existent.
One of the things we have to keep in mind is that these tools are not demonstrated to be very effective. Not just in terms of risking pulling over the wrong person, studies have shown that these fail to deter things like car thefts, they don’t provide a general crime deterrence. I did a study in—it was the Piedmont PD, which is a small city surrounded by Oakland. I did a study of their cameras because they kept basically the most complete data in the nation as far as I’m aware of. Their ALPR scans don’t correlate even with stolen vehicle recoveries. It doesn’t correlate with investigative leads, which is kind of—investigative leads could include things like identifying a suspect, locating a witness, spotting a stolen vehicle. They literally don’t help with that. So I’m not saying that it’s time to completely write these off, but as of right now, we need to do it right or not do it at all, and right now we’re not doing it right.
Beyond Stolen Cars: Intimate Life Details [27:06]
Bob Zadek: Now, the technology is most useful, if at all, in locating stolen vehicles, in finding criminals or suspected criminals who are escaping and they’re escaping by car. With widespread use of this technology, sooner or later they will trip a camera and the camera will help law enforcement locate them. Sounds like a good idea. But you have described a technology which goes well, well, well beyond finding the car. But we started our conversation with discussing what it tells us about the life of the user of the car.
So let’s leave the safe zone of finding a stolen car. Kind of works okay, you’ll find some—a fair amount of stolen cars. And you’ll find a couple of nerdy guys in the East Bay who were doing nothing wrong and found themselves on the evening news. But okay, you are collateral damage, Jonathan. Welcome to the club. Tell us about the misuse or the possible misuse of that information and what you suggest ought to be done with it. It’s the weaving together mosaic, and perhaps you can explain the concept of mosaic because that works its way into your article and it works its way into the conversation. So tell us what else as a byproduct this technology provides to anybody who wants it and what the dangers are.
Jonathan Hofer: Yeah, and I might characterize what we mentioned previously on just pulling over the wrong people and how it shows a picture of people’s lives. I kind of think of it as ALPRs are bad when they work incorrectly, and they’re probably even worse if they work perfectly. So that’s kind of what we’re talking about is their role in mass surveillance.
So a couple of things to consider is how the data is stored. And we touched on this briefly, but you know, sometimes it’s a common pool of information, but usually individual officers or departments will have access to the data. And you would think that given the sensitivity of this data, it would be tightly regulated or controlled or there would be a lot of safety features. Unfortunately, that’s usually not the case. What we found is that many of these departments aren’t even—they don’t have really good cybersecurity standards, sometimes they’ll give access to the wrong people. In some cases, officers retire from the department and they still have access to this database.
Sometimes you have nefarious bad actors who are using the databases to either stalk ex-wives or ex-girlfriends. There’s a kind of semi-notable case in Washington where other police officers were stalking this female police officer; they accessed her record something like 400 times. And then you also have to consider—one of the things that I think is worthwhile to mention is California, for example, has sanctuary city laws, and a part of that is not handing over data to, let’s say, Immigration and Customs Enforcement. Which is kind of a Jeffersonian principle, not to say anything about sanctuary cities, but there is something to be said of the people are explicitly saying that their local government shouldn’t be tasked with enforcing federal law. At the very least, it could be diverting valuable law enforcement resources. What data-sharing arrangements do is they skirt those laws. So let’s say you have loose privacy regulations in one jurisdiction. Well, you can use that jurisdiction’s database access to get data from that place, or you can say, “Hey ICE, here’s access to all our license plates. You know, we’re not helping enforce immigration, but here you go.”
Mosaic Theory and Legal Precedents [31:40]
Jonathan Hofer: And to your point about mosaic theory: so we briefly talked about how taking a picture of someone’s license plate is obviously not illegal. And I don’t mean to say that that in of itself is per se wrong. However, it does kind of encroach on unconstitutionality when you start putting pieces together. And this theory is what legal scholars call mosaic theory. And it says that even if collecting individual data points is legal, once you create a picture with those points, then it becomes a Fourth Amendment search.
And this chiefly stems from two semi-recent Supreme Court cases. In 2012, there is United States v. Jones, and I believe Scalia wrote the opinion on this. And I believe it was in the greater DC area, police try and put a GPS tracker on this guy’s car saying that he’s dealing drugs. This car travels out of the jurisdiction, the warrants expired, they’re still tracking him. Supreme Court unanimously said the government cannot GPS track a vehicle. Okay, then you flash forward a couple of years into 2017, Gorsuch is on the bench now, and there’s this Carpenter v. United States case. And I think this has to—the facts of the case revolve around I believe a bank robbery. And in that case, the government used cell phone tower data to track the cell phones of the suspects. Supreme Court said you can’t do that. Not only can you not do that, you can’t even retrace the steps of the suspect using historical data that exceeds 12 days.
Now, when you combine those two cases, I think you have a serious argument that says license plate readers as they are practiced in California and throughout the nation are unconstitutional. ALPR location data is not technically GPS, it’s not using a satellite. However, it is an effective substitute for GPS data. It’s logging your latitude and longitude even though it’s over the internet. And so they are creating a picture of your lives and they can use the historical data to track your whereabouts. So this has not been in front of the Supreme Court yet, and in fact there’s really no ALPR case that’s been before the Supreme Court. There’s really nothing that’s coming up that the Supreme Court’s going to rule on ALPR use generally. However, in lower states they have endorsed this mosaic theory. The most recent one that I’m aware of I think is Commonwealth v. McCarthy, and that’s the Massachusetts State Supreme Court. A woman was spotted I believe crossing a bridge where there was an ALPR over there. And there’s another one in the Texas maybe Court of Appeals—I’m not sure which district it was. They also endorsed the theory. However, in both those cases, the courts ruled in favor of the state. And their reasoning was that with only one camera and with only one scan in the Texas case, there wasn’t sufficient data to form a mosaic.
However, that’s not what’s happening in California. Piedmont, the city I mentioned previously, it’s less than two square miles and they have almost 40 cameras. In Los Angeles, there’s 500 cameras. These litter the streets. They’re scanning your car more than once. Some guy in San Leandro did a public records request on his car; it was scanned, you know, like 100 times. In my contention, that does warrant—that does constitute a mosaic search and therefore would require a warrant.
Private Collection and Civil Litigation [35:55]
Bob Zadek: So the danger is an unintended use because the technology is not sold as it will say—explain to anybody who cares where you and I have been, or at least where our car has been, and presumably we were in it, but that’s a different element of proof. But at least where our car has been over a period of time. Those who market the technology and those who advocate for its use are typically, as I appreciate it, law enforcement who just say, “We want to catch escaping felons and stolen cars and things of that nature, and perhaps Amber Alerts maybe.” They don’t sell it as “We’ll build the life, the hidden story of Jonathan Hofer so that you have no secrets, Jonathan, at least not anymore.” So I guess that’s a byproduct. That’s what I said earlier. That’s, “Hey, we have the data. How else can we use it?” And we are in law enforcement, so let’s look for law enforcement type uses like proving he’s a drug dealer as opposed to proving whether his car is stolen.
So I think what’s happened is once you have the technology, now that’s information that’s nowhere else available, so hey, why not use it for a different purpose? And then we have, as we always do when you mentioned several times decisions of the Supreme Court, the technology is always ahead of the law. And the law is trying to catch up. “Wait a minute, how do we apply a statute that was drafted when this technology didn’t exist? How do we apply it or not?” So that’s what I think you are explaining.
Now, there’s another question that occurred to me. And Jonathan, I’m going to ask you a question on the air to be safe for all time. But I say to myself, this guy is never going to know the answer to this question. So I tell the audience, when you don’t, it doesn’t mean you’re now humiliated. But what occurred to me is I was wondering, let’s say there is a civil dispute, a plain old boring lawsuit. And an element of proof that one side needs is the story which would be told by the mosaic theory on where somebody else’s car has been over a period of time to prove a point in a civil case. Now, if one of the parties in the civil case had a witness and could testify—maybe Jonathan, you had no life and all you did was you were obsessing about this car and you followed it all over the place, which you’re allowed to do, you weren’t stalking, that’s just your hobby, you follow this green car—and if you were to testify, you’re allowed to testify, there’s no privacy rights at all. Now, take the same need to prove and the litigant contacts the private company that maintains these cameras and says, “Hey, we understand you got these cameras all around town. Can we—can we pay you? We want to run a license plate search and want to know where this car has been for the last four weeks.” There’s nothing wrong with that. That doesn’t offend you, or does it?
Jonathan Hofer: Well, that’s a great point. So that actually has come up in a court case, interestingly enough. So the argument was—and this was from a company that aggregates data and leases it or rents it out to police—and they said, “We have a First Amendment right to this data.” And they asked the question, since license plates are in public, what’s the difference between us and a newspaper? Because a newspaper just, you know, they use their eyeballs, they observe the facts, they record it and publish it. What’s the difference? I think that that’s a very important argument to address. Unfortunately, there’s no court case where this is butted up against mosaic theory. And I would hold that, you know, private people and companies do have a right to store this data. However, it does become concerning once the government is using this as effectively mass surveillance.
Bob Zadek: So isn’t there—so if all the data, if none of the data was collected by government, was all collected by private parties, and maybe each county franchised it out to one license plate reader collection company, but now we have the data lawfully being collected because you told me that’s allowed by private parties. And now it’s like cell phone records collected by Verizon, and the police need a warrant to get that cell phone data from Verizon, and they get the warrant if they can satisfy a judge they’re entitled to it. Would that system, putting aside possibility for abuse—there’s always abuse—would your concern be materially satisfied if we simply said law enforcement could never be the collector?
Jonathan Hofer: I don’t necessarily have a problem with law enforcement collecting their own scans. And I would say that law enforcement requiring a warrant to get it from third parties, I think would be eons ahead of where we are now. So I think most of my concerns would actually be absolved. And honestly, I don’t have a problem with private companies—like toll collection or things like parking enforcement in private parking garages, I am totally okay with that. I don’t think the government has a vested interest in keeping records on people not suspected of criminal wrongdoing. And so while I don’t criticize private individuals or companies collecting license plate records, I do have a problem with what the government would do with those.
Remedies and the Future of Surveillance [43:00]
Bob Zadek: Is there any kind of—because clearly there is the possibility of abuse, and clearly there are examples, you being one of them. And I didn’t realize I found this goldmine perfect guest on my show. I had no idea you not only talk the talk, you walk the walk, but you found yourself lying face down in the pavement at a dark parking lot somewhere in the East Bay. So how lucky we all are to have you on the show. Would your concerns be mitigated, maybe mitigated substantially, if there was a system that adequately—air quotes—adequately compensated the victims? And now you have the cost of the system. When governments decide whether to buy it, the cost to the system would be the total of what the government pays for the system, for the hardware and the maintenance, plus the error factor and the cost of that error. So that the victims of mistakes get compensated. Works so they’re as best they can be made whole—you couldn’t be, how do you compensate you for lying face down in the pavement when you just want to be home under the sheets? But so you figure out a way to do that. That’s for somebody else to do. And if you compensate the victims and the cost is borne by the source of the mistake, as a libertarian or libertarian-ish, I will presume libertarian, are you okay now? Is the downside covered so now we only get the upside and society as a whole is not penalized?
Jonathan Hofer: Yeah, I mean, in that case, we really want to make sure that the benefits of the technology outweigh all possible costs. So not just, you know, lawsuits, but also any maintenance and other errors this caused. I don’t want to fleece the taxpayers with the burden of law enforcement incorrectly pulling over people. However, I do think it’s prudent that cities that do operate ALPRs, they carve out some remedy for the people they’ve wronged. San Francisco has done this in their license plate reader policy; they’ve carved out a private right of action. They say, “If our surveillance equipment injured you, you have the right to sue us if we don’t make it better.” I think that is important to do. You know, maybe a broader thing would—that also brings up the issue of like, do police have malpractice insurance? You know, do they have a stockpile of this that they can pay out lawsuits? That brings up a lot of other issues.
Bob Zadek: We’re getting into qualified immunity, aren’t we?
Jonathan Hofer: Yeah, yeah. So that—regretfully, I would love to do it, that’s a whole other show, a show that I have done and will do more of.
Facial Recognition and the Road Ahead [46:16]
Bob Zadek: Now, as we start to run out of time, one last thought, Jonathan, briefly. This conversation could just as well be on facial recognition. We have CCR cameras all over the place. The UK has gazillions of them. Not only do we know where cars are, we know where your face has been, and presumably you’re attached to it, so therefore we know where you are. How much of what we have discussed today apply with equal force to facial recognition, and is one more sinister than the other? We only have about a minute or two, Jonathan.
Jonathan Hofer: Yeah, I’ll be quick. Yes, the issues are very similar. In fact, they’re used similarly tactically by police. They have the same data collection problems, the same tracing someone’s steps problems. I think the bigger concern is actually license plate readers just based on their number. Facial recognition’s kind of far and few between, and a lot of cities have outright banned facial recognition like Seattle, Portland, San Francisco, a few others on the East Coast. Most cities have not banned license plate readers; they’re used everywhere.
Bob Zadek: Jonathan, on the subject of facial recognition, now, I can’t figure out whether it’s easier to change your face so the cameras don’t get you or your license plate. I don’t know which is harder. But putting that issue aside, what is the future? We now have, of course, where we are with facial recognition and with license plate readers. What’s the next step in that technology, just so nobody goes to bed tonight feeling private?
Jonathan Hofer: Yeah, one of my big fears is that the United States will quickly adopt a lot of the tactics and strategies that the CCP uses to control its Uyghur population. The Chinese government has used a lot of surveillance tools, most notably facial recognition. They have large databases on people’s purchases, internet search history. United States already has internet search history on lockdown with the NSA. And few people realize that the United States is one of the largest importers of Chinese surveillance technology in the world. However, I think we have reason to also be optimistic. The story’s not over and there’s a lot of things that people could do. There’s a lot of votes that have passed in the Bay Area and elsewhere where we’ve successfully curbed government encroachment into people’s privacy, and I think that could be replicated elsewhere.
Bob Zadek: The optimistic point of view is I don’t think we’re going to stop the technology. Therefore, the issue is not whether the government has or even uses technology to collect. The issue is the use. And it’s easy to control the use if you misuse it, you go to jail or you pay a fine. And that’s the way to control the use. So rather than try to either stifle the technology or control the technology—that’s not going to happen—let’s just control the use and limit what government is allowed to do with the technology. I find myself not caring all that much if somebody in Tulsa, Oklahoma, who will never meet me, knows that I went to have donuts yesterday, as long as she doesn’t tell my wife. I don’t care if she knows that I had donuts. But if she misuses it in her public capacity, that’s what gets me concerned. And I think the prospect of controlling the use, while difficult, there’s reason to be optimistic, but there’s no prospect to control the technology.
So Jonathan, if we both can agree on that, we leave the show neither one of us is going to be accused of being a Luddite. We can walk proudly in libertarian comfort and not be ashamed. Tell us about the Independent Institute, what they are, what they do, and how folks can follow your work and the work of the Institute.
Jonathan Hofer: Yeah, Independent Institute is a public policy think tank located in Oakland, California. Our website is independent.org and our mission is to advance free people and free markets.
Bob Zadek: We’ve been talking to Jonathan Hofer. Jonathan is a research associate at the Independent Institute located in Oakland, California. They are a lead—they have a huge family of scholars and experts that publish regularly under the Independent Institute, as does Jonathan. Jonathan, do you run an active Twitter feed or column where the audience can follow your work?
Jonathan Hofer: Yeah, the best way to follow my work is searching up my profile on independent.org. I have my own author profile, you can read commentary, blog posts, things of that nature.
Bob Zadek: That’s Jonathan Hofer, H-O-F-E-R. Jonathan, thank you so much. I’m sorry about your loss when you were lying face down in Oakland. If I went to that parking lot, would I see an outline in chalk of your body lying there? Did they save it as like a monument?
Jonathan Hofer: I hope they did.
Bob Zadek: You’re famous, Jonathan. Jonathan, thank you so much for your time. And to all my friends out there, thank you for giving us an hour of your most valuable time. We hope you have found it worthwhile. So long for now.