See Tracy Rosenberg at this year’s event! She will be on a panel with Lisa Rein and Daniel Rigmaiden on Saturday at this year’s San Francisco hackathon!
TICKETS ($50) – (Can’t afford it? Okay no problem. Just use the code “hackathoner” for 1/2 off! (Perhaps you can buy one at full price and then discount the rest? Or just use the discount for all of them! It’s up to you 🙂 – Yep – It’s on the honor system folks <3
|Saturday Schedule & Evening Event||Projects To Hack On|
This is the second of two interviews with Tracy. Here is the first interview, which is about how Oakland Privacy’s roots are based in the “Occupy Oakland Privacy Working Group”.
By Lisa Rein.
The Aaron Swartz Day Police Surveillance Project is all about developing a larger strategy for determining what police and sheriff departments have already purchased, and putting “surveillance policy frameworks” in place to monitor and regulate the use of that equipment against their residents.
The project first started during Aaron Swartz Day 2017’s Sunday hackathon. Before that event was even over, it was clear that it had been really successful and we were all very pumped and decided to just keep going on our projects, through till next year’s event.
Aaron Swartz filed a lot of FOIA requests, and it made me want to start a project at the hackathon that would continue the tradition. (Note: FOIAs are “Freedom of Information Act Requests – and are Federal. This article discusses “public records requests” at the local, or municipal level.)
Tracy has been teaching me about how powerful the information obtained from public records requests can be, because it is literally the police or sheriff department saying, themselves, “Yup. We have this piece of equipment. Here is our purchase receipt.” (For example.)
However, writing the letters and filing the requests can be time consuming, even with systems like Muckrock, that vastly improve the process. So, our first idea was to automate the process of using Muckrock, so someone could fill out a form that would connect to Muckrock seamlessly. It turned out though, that except for the automation, Muckrock already had everything we needed. So, we decided to use Muckrock, and focus on saving folks time by providing all of the letters ahead of time for every single piece of equipment (including requesting details about how the data from these devices was collected and stored).
Tracy wrote up all the letters ahead of time, turning the job of writing and submitting 18 public records requests into a quick cut and paste job. Together, we filed together 18 public records requests; 9 to the City of Sacramento and 9 to the county of Sacramento, in just under over two hours, but it was our first time. (Here’s what we found in Sacramento.) – Here are our letter templates.)
In the meantime, I took a few moments to talk to Tracy about the implications of filing public records requests, and how the information obtained from them eventually led to a first of its kind Privacy Commission and surveillance policy framework, in Oakland, California.
Lisa Rein: First, let’s talk about the kinds of information we can uncover with public records requests
Tracy Rosenberg: We get a pretty clear example of the kinds of things that we can find out using this latest find about license plate readers as an example.
LR: So first, please explain to me how these license plate readers work.
TR: Sure. A little about license plate readers, for people who don’t know what they are. These are cameras that are set up, either on top of police cars, or at fixed locations. These are often located on highways or on major intersections. And what they do is take pictures of the front of every car. Every single one that drives past.
So you get this monumental file, every day, of all the traffic in a given location. And of course these license plates can be linked to the owners of the cars in the DMV’s databases. So it basically provides geolocation data for an individual, and if you happen to go by one of these license plate readers every single day, because it’s in your town, it really provides a record of all of the comings and goings, for you as an individual and for the inhabitants of your car.
One example of how license plate readers can be misused, is when the New York Police Department sets them up in the parking lot of mosques, in order to get the license plate numbers of people who are attending services there. Although the act of attending services at a mosque does not itself make you a criminal.
LR: So in that case, it’s clearly being misused to conduct a form of racial profiling?
TR: Yes. It’s a form of racial profiling against a specific population.
LR: It seems that these civil rights issues should be non-partisan in nature: being innocent until proven guilty, the importance of probable cause, and requiring a warrant to be issued before records from these systems can be accessed legally.
TR: Yes, I agree. There are libertarians and chunks of the right that also have concerns about this sort of wholesale checking of data. In Sacramento, a Republican/Libertarian senator brought up a bill about restricting license plate readers, because he was concerned they had license plate readers set up in Southern California at a gun show. He thought that was a huge violation of privacy and civil rights.
Perhaps if Homeland Security stood on the corner and looked at everyone’s papers before we crossed the street, they might catch a few more bad people. The problem with that is that we would be living in a police state. So, we have to draw the line somewhere about what is and what is not appropriate. Because, if we go all the way to “everything is ok, if we catch a few more bad people,” then the rest of us are going to be living in hell. There is always a line between personal freedom and security, and where that balance is is really a community conversation. It’s not just a law enforcement conversation.
LR: That’s another goal behind filing these public records requests and using the information you get to take a stand as a community. That way, at least we can have some kind of a voice in the official process dealing with this stuff.
TR: Absolutely. And to get outside of partisanship, that community voice might sound a little different in Bakersfield that it does in Oakland. That’s okay. What matters is that affected community gets to speak up.
LR: Okay. Back to the license plate readers.
TR: So, Dave Maass at the EFF, who is also on our team for this project, filed some public records requests with The City of Sacramento’s Police Department, asking them who they were sharing their license plate reader data with. And when the Sacramento Police Department finally replied, we found out they were making their license plate data available to 792 different law enforcement agencies across the country. And we are talking all of it. Years worth. From a couple of dozen readers in the City of Sacramento.
LR: Are all 792 of those law enforcement agencies?
TR: No. Actually, some of them weren’t even law enforcement agencies. It included the United States Post Office. It included the Franchise Tax Board. It included the Yellowstone National Park.
It included other police departments across the U.S. Cities as obscure as Zebulon, North Carolina. Greece, New York. Places that make you wonder: “why do they need to know about people driving through Sacramento?” I mean what was the law enforcement purpose here? It also included the DEA Chicago Strike Force Team and three Homeland Security ICE offices, including Chicago and New York, and Dallas. It included the FBI.
Some of it’s surprising. Some of it’s not. But Smokey the Bear? Why?
LR: So, rather than a system where anyone who asks can see the data, it would appear that anyone on the system can already access all of the data?
TR: Yes. Anyone can access the system if “sharing” is turned on. We know now that any of these 792 agencies can get into the City of Sacramento’s database and take a look, whenever they want.
LR: Because Sacramento has “shared” its data?
TR: Yes. They have shared it. They have not done anything to protect the privacy of people that just happened to be driving through Sacramento, who have not done anything wrong, and are not under suspicion of anything.
Remember that the whole issue with mass surveillance is that it’s not connected to any kind of reasonable suspicion, probable cause, or doing anything wrong. It’s just everybody, and essentially the information is collected in order to have it. It’s the opposite of what law enforcement calls “need to know,” which is that information should be made available if there is a law enforcement reason for it. But really, what they’re doing is going out and collecting it just because they can.
LR:. So the shocking thing here is not only the extent of information collection on a population that is not yet under investigation for anything. But also, the extent of information sharing that is largely unknown and largely unregulated.
TR: Yes, because there are often no policies put in place. So the problems range from 1) Where is this happening? 2) What’s being collected? 3) How much is being collected & over what time period is it stored? 4) Where is the data being kept and is it being kept securely? 5) Who can see it and who has access to it.
We want some sort of control over this process, and in a democratic system that comes via our representatives. It becomes a question of your elected officials representing you, and of grassroots efforts putting pressure on them to take these issues seriously. But they can’t get the chance to do that if they don’t even know what’s going on, or if they don’t have any ability to regulate or place any restrictions and limits on what’s going on, once they do know.
LR: So that’s where these official statements about the equipment being purchased and implemented, coming straight from the source, make a difference, right? And this is the point of the Police Surveillance Project; getting the evidence, so you can say to your representative “Hey, the cops admit that this exists.”
TR: Right. I know for a fact, that if I went to the Sacramento City Council, and told them, prior to having this piece of paper in my hands, that they were sharing license plate data with 792 other agencies, they would tell me that I was full of crap. This way, we both know it’s a fact that has to be dealt with. We can skip the debate about its existence, and get on to the important questions.
LR: So! Now we have the evidence, and we know that it’s happening, what can we do about it? So, the idea is to say “Hey, if this equipment is going to be kept at all, let’s at least get a surveillance policy framework firmly in place, with clear limits on how it is allowed to be used on citizens and how any data it collects is allowed to be shared.”
TR: Right. In this case, do these 792 agencies really need to know where everybody in Sacramento drives, and if so, why? And also asking “do the people in those cars have any right to weigh in on whether they are comfortable having their data shared? — as I think they do. But you need to be able to have the data in your hands first for the people involved in the political process to be able to acknowledge your point of view.
LR: So now that we kind of have the “smoking gun” for license plate readers in Sacramento, what do we do next? Who do we talk to about this information that we now have?
TR: Well when it comes to Sacramento, and I just wanna get away from that for one minute, to talk about the fact that sometimes it’s not just a question of how it’s being used and how it’s being shared, but a question of “is it happening at all?”
We have an example of that from the City of New Orleans, which is not in California, where it just came out on a FOIA that they had been using predictive policing software developed by Palantir, the CIA funded Peter Thiel company. And the city council didn’t know, even though they’ve been doing it for years. They didn’t even know they had the software program at all. They didn’t know this was happening. Not one little bit.
So, in addition to the equipment that is just unregulated, has no policy in place, or is going further than it should, there is also secret surveillance going on that essentially the political system doesn’t even know about. It’s happening behind their backs, and of course, if it’s happening behind the city council’s back, it’s happening behind the peoples’ backs.
LR: But this doesn’t have to be an adversarial relationship between us and the council members. This can be “fact finding” that provides them with better information, so that they can make more informed decisions. Right?
TR: Yes. We often find that they actually want to know. I’m pretty sure if or when I go to Sacramento and say “You are sharing your license plate data with 792 agencies. Did you know you were sending it to Yellowstone National Park?” They are going to say “of course not.” Because they didn’t know, and I’m sure that they didn’t know.
Then, the question becomes: “Well, now that your constituents know that you’re sharing your data with Yellowstone Park, and the Franchise Tax Board, and the DEA, 1) you should ask them what they think about that and 2) what are you gonna do about it. And that’s when they say “what can we do about it?”
And there is the moment of truth, where, as advocates, we can say…. “policies!”
Transparency policies work best when there is a whole framework of policies in place, so that all of these surveillance issues come to the government first, so they get a chance to say “yes” or “no” to whatever it is.
LR: So that provides a catch all system for whatever new surveillance hardware or software comes across the table.
TR: Right. So, for example, “Yes, we are going to have license plate readers. Here is where we are going to store the data. Here’s how long we are going to keep it. Here’s who has access to it. Here’s who doesn’t have access to it.” You make an affirmative, purposeful, decision, all in public, that people can see.
Maybe the license plate readers go away. For some opposed to mass surveillance like me, that would be nice. But if they don’t vanish, then at least we know what’s happening, and can be sure there are some limits in place. If they are going to share the data with 792 agencies, we at least know that, and we won’t have to do a public records request a year after the fact to find out. (Although it’s important to keep filing public records requests to check up on the usage reports from law enforcement.)
Also by making the people you voted for weigh in on those decisions, you can now hold them accountable. My guess is that if they have to do it in public, then they’re not going to advocate sharing it with 792 agencies, because that’s absurd and that’s not what most people want.
LR: Okay so let’s talk about the City of Oakland, which has become nothing less than a model city in the way that it has implemented this kind of ideal “surveillance policy framework system” that you describe. Oakland Privacy worked with the City Council to get that implemented, yes?
TR: Yes. Oakland Privacy, the organization that I’m with, basically came about as a community response to a big huge Homeland Security project that was sort of secretly started in Oakland. (It wasn’t entirely secret, but they weren’t really talking about it that much.)
LR: Yes, I remember when that project (the DAC – Domain Awareness Center) got leaked on Twitter.
TR: Yes, it was a city-wide gauntlet of cameras and microphones and license plate readers on a citywide grid with a centralized fusion center. It would have turned the whole city into sort of a 1984 come true.
What we found out was that they had been sort of sneaking through on what they call the “Consent Calendar” at the City Council. So they sort of voted on it, but it went by really fast, it was over in a flash, and nobody really knew what was going on.
LR: So you guys stopped it?
TR: We did stop it, and we did it largely by sunlighting it, so that everybody said “what the heck?” And the City Council that had been just shoving it through started to step back a little and say “well, we didn’t realize it was that big. We didn’t realize it was that comprehensive. Maybe we should think about this a little more.” And eventually, they walked away from it.
LR: It had this whole “nick of time” feel to it though, because they were going to vote on it.
TR: It was already up to Phase Two. So our representatives that backed away from it, didn’t start off backing away from it. But when we surfaced it, and they had to do it in public, it changed. So, as a result of this successful community mobilization against the DAC – (the Domain Awareness Center), Oakland has gone on a different trajectory than any other city in California, and voluntarily set up a citywide Privacy Commission. The first one in the country. No law enforcement on it; just people.
LR: Wow. So, it’s basically a model city.
TR: Yes. The principles of surveillance transparency are in place in Oakland. The whole modus operandi around this Police Surveillance project, is to crowd source the massive project of convincing every other city in California that they need to do the same. That’s what we’re trying to do.
And in fact we now have a model law that can help you do that. It’s on the OaklandPrivacy.org website.
TR: The model legislation that Oakland Privacy and the ACLU and Restore the Fourth and a few other groups in the bay area have created, creates a framework. So, rather than having to pass a law for every piece of equipment, which is a problem because there’s always new devices. We have a framework that says “every piece of equipment. No matter what it is – if it performs some kind of surveillance function – it comes under this framework.
LR: So the framework sort of future-proofs the whole thing.
TR: Right. Now they can invent whatever new device they want – which they will – and it will still come under this framework.
LR: What does the framework say basically?
TR: It basically says “Law Enforcement, you have to tell us in really specific terms how this piece of equipment is going to be used. Either before you buy it, or, if you’ve already bought it, before you use it. That’s where the City Council can say “I don’t want it used at all at protests. Period.” Or “We don’t want it used without a warrant.” Which is what people are using today as sort of the new Stingray standard.
Saying “you don’t use it without a criminal warrant.” Here in Oakland, that particular requirement really changed the way that it was used, and we went from using it a lot to using it 3 times a year. So you went from something that would appall most people, to something that most people can live with. Although myself, I think three times could be three times too many, but it was used with a criminal warrant, and all three cases were for murder or attempted murder.
TR: So, to finish explaining how the policy works: Basically, you get this framework into place, and every piece of surveillance equipment has to go through this process. And then the other part of it is that every year they have to report back on what they actually did. For example, the Stingray, in Oakland, implemented a policy that said “used with a criminal warrant only” – and then, a year later, we got to find out “how many times did you use the Stingray, and did you have a criminal warrant every time?” They have to report back and give us the documentation. So, they reported back that they used it 3 times, and that they had a criminal warrant each time, and that the charges on the warrants were for “murder,” “murder,” and “attempted murder.”
LR: And this isn’t just the honor system, right? They have to give you written documentation?
TR: Yes. It’s a report to their City Council, and essentially the way that it works is that they are required to submit back up documentation. The specifics are held back from the public. The public sees the report, but not necessarily the documentation, due to the privacy issues that would involve. But the City Council has access to the backup information.
LR: And it is clear that the Police Representatives would be doing something illegal if any information were to be misrepresented in a report to the City Council?
TR: Yes, the way the ordinance is written, any kind of misrepresentation would be a “punishable offense.” The ordinance also gives a “private right of action” to any members of the public that gives us all standing to sue or intervene. We can take legal action if there’s no report or if there is misrepresentation or it is faulty in some way.
LR: And, so far so good, in Oakland?
TR: Yes. So far everything seems accurate and following the rules. But this is why it’s important for us to keep up with the public records requests.
LR: Oh. We need to keep doing them. As a fail safe?
TR: Yes. To cross reference and double check information. Look what happened in the City of New Orleans; they sent a reporter information, via a public records request, about predictive policing that hadn’t even been given to its City Council.
LR: No one from the City Council had asked for that information.
TR: No, but a reporter asked; and they handed it right over. You know, a lot of this stuff goes on in secret. Not because they are forceably concealing it, but because we are not asking the right questions.
LR: So, even after we have firm surveillance policies in place, we need to continue to file these public records requests, as a check and balance, on the accuracy of the reports by various law enforcement agencies given to their respective City Councils. Does it get any easier as time goes on?
TR: Well once you get the framework in place, then the policies will exist, and it will be much easier to check up in a targeted way.
LR: And it’s been working for the most part, here in Oakland, and for how long?
TR: We defeated the DAC, that big city-wide surveillance gauntlet, back in 2014. We got the Privacy Commission set up in 2015/2016, so we are really just coming around now to the first cycle.
We’ve gotten this on the books with the County of Santa Clara too and in the cities of Berkeley and Davis. Santa Clara did a very extensive set of reports. They reported on I think 40 different pieces of equipment. A really large number, and that’s a lot of information.
LR: So, so far so good.
TR: Yep. So far so good. We’ve seen some inhibitory impact from having to go through the process. A number of polices have been put into place, that were not there before, that have been strengthened, and we’ve seen less usage, because they have to report on it. Because these reports are public. Because people will see them. Because the City Council will see them, and because they can end up in the newspaper.
LR: So the numbers are going down every year? Since this started?
TR: Yes. We’re seeing some inhibitory impact. Certainly, in Oakland, when we put the warrant requirement on the Stingray, usage dropped. In Santa Clara, we also saw pretty limited use, and we have reasons to believe that, before we went through this process, the use was probably much more extensive.
LR: Do you think the word is getting out surveillance policies?
TR: Yes. I think that Law Enforcement is beginning to understand that, for everything they do, they’re going to have to have usage policies.
LR: Do you think there’s any chance that some departments may proactively cut down on their usage of some of this equipment, since they know they will have to report on it?
TR: Yes. I think so, and I think we will see that impact on other cities, as we go through this process.
LR: Well, thank you so much for helping us with the Aaron Swartz Day Police Surveillance Project, and for teaching us all about these public records requests. Anything else to mention?
TR: Well I’d just like to encourage people and let them know that’s it’s really easy to make these public records requests. Once you get the hang of it, you literally will get all 18 done for your city in a couple hours. And the information that you get back will really be invaluable.
LR: Yes, just to explain. Tracy has already done all the work here. She wrote up 18 templates – covering every type of surveillance equipment/records collected as possible. All you need to do is cut and paste your name and mailing address into the templates, and then use Muckrock to find the Police and Sheriff agencies in your local area. I’ll have a tutorial up very soon.
The cut and paste into Muckrock is all very very simple. Tracy and I filed the first 9 requests in about 1.5 hours and she did the the second 9 herself in about 45 minutes. (Reminder to let me know if you decide to do a city, at lisa @ lisarein.com – so we don’t double up too much, since we are trying to get all the major cities in the state.)
TR: We definitely want to have data from every assembly district in California. It’s something like 90 of them.
LR: Right! See the project page on the Aaron Swartz Day Website for more information.