Get Informed

Subscribe to our newsletters for regular updates, analysis and context straight to your email.

Close Newsletter Signup

The Appeal Podcast Episode 5: How Silicon Valley Supports LAPD Surveillance

With Appeal staff reporter George Joseph.

LAPD officers line up in front of protesters.
Lucy Nicholson / Getty

The Appeal Podcast Episode 5: How Silicon Valley Supports LAPD Surveillance

With Appeal staff reporter George Joseph.

Tech companies like Palantir are teaming up with police departments to create real time “probable offender” lists of thousands of mostly African American and Latino people. These lists, and other “predictive policing” tools, create a feedback loop, trapping people of color and the poor in a cycle of monitoring, arrests, and further monitoring. The result is a dystopian system of surveillance and pre-crime. Our guest, Appeal journalist George Joseph, got his hands on confidential LAPD documents detailing these efforts.

The Appeal is available on iTunesSoundcloud and LibSyn RSS. You can also check us out on Twitter.


Adam Johnson: Welcome to The Appeal Podcast. I’m your host Adam Johnson. The Appeal is a podcast on criminal justice reform, abolition and everything in between. You can follow us on Twitter @TheAppealPod and on Facebook at The Appeal Podcast. Large corporations like Palantir, are increasingly teaming up with police departments to create what is effectively a real time rap sheet on thousands of mostly African American and Latino people throughout the country. Our guest today, George Joseph, a writer at, he got his hands on one set of these documents by the LAPD that was made in concert with Palantir. These documents are a fascinating window into the truly creepy nature of these new pre-crime tools, what they mean for privacy and due process and we’re going to have him on to talk about this and how activists are fighting back against these forces.

[Begin Clip]

George Joseph: When police are sort of being tasked to just go after specific individuals because they happen to be on a list, even though there’s no sort of reasonable suspicion or probable cause against them, it really creates the environment for those who are targeted of a kind of police state where everywhere they go, every interaction they have with someone they know in their community is being potentially documented as being potentially to watch even if they don’t know it’s being watched.

[End Clip]

Adam: George, thank you so much for joining us on The Appeal. I really appreciate it.

George Joseph: Hey, thanks for having me Adam.

[Fade Music]

Adam: Uh, so you wrote a fascinating piece in May called “The LAPD Has a New Surveillance Formula Powered by Palantir,” which is an incredibly neutral and sober headline, but really I think in many ways downplays actually how pernicious this really is and I think when we talk a lot about new surveillance technologies we’ve almost become kind of immune to them. They sort of, oh well of course we’re being surveilled, but there’s tons of, of really dangerous stuff here that I want to get into and I want to talk about the broader trend of using what is effectively pre-crime as a tool to reinforce racist and anti poor policing regimes with, with the kind of new techno veneer. So can you start off by talking to us about what your article is covering, specifically the LAPD’s Chronic Offender Bulletins and their partnership with Palantir, can you sort of set the table for people?

George Joseph: Sure. So the LAPD has had a program known as the Chronic Offender Bulletin program, um, for several years, in fact, before Palantir software even came on board to the LAPD. What the basic idea of the program is, is that in select geographic zones, intelligence analysts in the department will look at individuals who have lengthy criminal histories and other kinds of factors and then kind of create a pile of sort of top target priorities for police. Now they claim that this isn’t being done, um, you know, to explicitly arrest them. But the documents that we found and were given by a coalition known as the Stop LAPD Spying Coalition, show that they’re very clearly being, um, sort of judged based on how many arrests of these kinds of individuals they’ve made. Even though the documents make clear these people aren’t actually wanted with any probable cause for arrest. Its just sort of setting up a system in which people get arrested because they’re somehow deemed priority arrests without any probable cause or reasonable suspicion beforehand.

Adam: There’s a series of criteria, correct, that are incredibly sort of vague and if you meet certain amount of points on that criteria, you get on this, this mystery list, correct?

George Joseph: Right. There’s actually a checklist that analysts are supposed to check off to decide who has the most points and thus who is the most quote unquote “probable offenders” who should be gone after. What’s new about our article is that we were the first to report on the newest updates of that checklist, which for example, enhance some of the penalties about guns that were in the previous version of the checklist. So now it’s not just having a handgun found on you counts against you, but also any other type of gun. So we’re seeing sort of the gun criminalization panic effecting people in this very day-to-day way. Um, in addition, there’s other categories that were already there, um, from previous versions and still are there, for example, how many times you’ve been stopped by police count against you in this formula. So you are considered more high risk and thus more, um, sort of prone to surveillance if police have stopped you a bunch of times. But can you really control something like that? Obviously not.

Adam: Yeah. And this is what activists have called a quote, “racist feedback loop”

George Joseph: Right.

Adam: Now, this may seem obvious to some, but for those who it’s not, this is the way people, quote unquote, you know, “interact” with police and it’s typically based on where they live because police surveil certain parts of the city more than others. Even if you adjust for crime rates it’s almost always disproportionately places where it’s people of color, Latino and black and they’re far more likely to have interactions with the police. And thus far more likely to be on this list and thus have further surveillance and then have more interaction with the police. Um, until you go down this, this feedback loop of what is basically just a kind of harassment regime.

George Joseph: Right. And it also creates this oversampling problem where they’ll say, ‘Look, this isn’t, you know, racially biased, this is neutral,’ but you start with the initial sample, which is already based on, as activists have called, kind of biased data in terms of where police are policing, who they’re stopping, what are your obviously poor black people mostly and also Latino people sometimes. Um, but then when that sample is created, a risk assessment formula is based on that data. So that’s just the first round because then that data is collected based on those standards which have already been set based on the previous, uh, biased data. And so what you’ll find, you’ll have this oversampling problem where you’re only looking for those kind of people again, um, and you’ll find even more of them because you’re increasingly not doing kind of randomized policing.

Adam: Right.

George Joseph: And so every iteration of the risk formula will be more targeted and more biased against very specific kinds of communities. Um, so we see this ratcheting effect that has real impact on people’s lives.

Adam: Yeah, I mean anyone who’s had the misfortune of living next to a frat house where you have a bunch of 18 year olds getting wasted and jumping off the roof and a series of sexual assaults that go completely ignored by the police, knows that where the police target these, these alleged crimes is obviously deeply racialized and they try to go for the lowest hanging fruit. Um, and it seems like in many ways this is compounding this problem. There’s, there’s one quote you have from someone who was in a focus group that was commissioned by the Stop LAPD Spying Coalition for, for a report that was released in May. He said, quote, “I feel like they already know who you are by the time they stop you and give you a citation. They already know your name and who you are hanging out with.” This is obviously very dystopian stuff. Can we talk about the psychological impact it’s had on those under its charge?

George Joseph: So when police are sort of being tasked to just go after specific individuals because they happened to be on a list, even though there’s no sort of reasonable suspicion or probable cause against them, it, it really creates the environment for those who are targeted of a kind of police state where everywhere they go, every interaction they have with someone they know in their community is being potentially documented as being potentially watched, even if they don’t know it’s being watched and so people pick up on that and it, uh, affects how they feel they can move around in the world. Um, it affects the degree to which they basically just feel freedom living in day-to-day life. Um, when we talk about sort of authoritarian regimes and other countries as US citizens, we often don’t think about how those same kind of intensive day-to-day surveillance tactics are being put on people here because it’s only going to a very targeted and racialized slice of the population.

Adam: Let’s talk about Palantir, the Silicon Valley startup that was behind, that provides the technology. Palantir has been in lots of stories involving police surveillance, government surveillance, a lot of military application. Now Palantir as you know was started by, paradoxically of course, a libertarian billionaire. I’m not quite sure what that term means. In 2003, you did mention though that this is, I think somewhat interesting is that of course Palantir was also created by the Central Intelligence Agency.

George Joseph: Yeah. They invested in it.

Adam: They were the first investor through In-Q-Tel, and then it sort of went into the shadows for a while and then kind of reemerged in 2005 as a full blown surveillance company that deals primarily with things like monitoring social media and doing data analysis for these kinds of surveillance technology. Can we talk about Palantir and what other cookie jars they have their hands in vis-à-vis pre-crime?

George Joseph: Yeah, so Palantir gets a lot more attention than other kinds of surveillance software firms because it has this very kind of shadowy history of CIA funding and Peter Thiel who does weird stuff with blood and stuff like that.

Adam: Right.

George Joseph: Um, but in reality, what their software’s doing isn’t actually very complicated, um, or even necessarily that sophisticated. What they’re doing is being good at cleaning up in connecting lots of very disparate sources of data, most of which police are already collecting. But I, I don’t mean to underestimate how powerful that is because, you know, 20 years ago if a name was just in a file cabinet somewhere and a gun was in another precinct cabinet somewhere else, those, that link wouldn’t have been necessarily made. Or if it were to be made, it would take a lot more effort and just hands on investigation. Now, as all of this data is being automatically uploaded into kind of readable formats, where interoperable databases are communicating with each other and linking the data points together about people’s homes, their associates, their guns, all kinds of objects associated with them, police can, with a few clicks, map out networks and communication patterns and location patterns that they never would’ve been able to, in the same amount of time as before. So it kind of is just making much more efficient, um, police practices that have existed for a long time but would never have been imaginable today. And there are other companies doing this as well.

Adam: Right. What are those companies?

George Joseph: IBM, for example, used to have a, a, a pretty well known program called Coplink, which they recently sold to another company called Forensic Logic. I think that’s considered one of the bigger rivals to Palantir’s data software. There’s other companies like Cellebrite, which specialize in doing cell phone extraction and then with the extraction of that data, doing some of the same kind of link analyses that we see with Palantir where they’ll tie together people’s contacts and sort of make maps of their communications networks and that kind of thing. So it’s not really the software that matters as much as the ease with which police can grab data and connect data, which is what Palantir does specifically for the LAPD with this program. And because of the mass of this data, it becomes pretty powerful.

Adam: So let’s back up here in, in the people you’ve spoken to, and I know you, I know you’ve done quite a bit of reporting in this space for years now, what are the legal ramifications if any, are these police departments pretty good at kind of avoiding overt First Amendment violations or other kinds of regulations? And to what extent is the current legal literature, as it were, or the laws, catching up to this technology?

George Joseph: That’s a good question. I’m not a legal expert on, on what laws govern what’s increasingly known as “predictive policing.” Um, but what we see is that police are very careful in their language about what they’re doing when they engage in these kind of programs. So for example, with the LAPD program, they’ll give officers a printout of an individual with all their information and their face and where they’ve been stopped recently and their history, of one of these so called priority or chronic offenders. But on the sheet it’ll say ‘This is only for informational and law enforcement safety purposes. We do not have probable cause to arrest this person.’

Adam: Right.

George Joseph: So they’re winking at their officers like ‘Just say no, this guy is of person of interest, but we can’t technically go after him unless you happen to see him doing something which we’re hoping you will do if you’re constantly following them around and eventually catch them doing something.’ And we know that that’s the implication because in the slides we published, the LAPD would for example, say things like, ‘Commander should be telling their supervisors how many people from the list have been taken off and taken into custody since the week started and analysts are expected to maintain a list of sort of a deck of 12 priority offenders and then constantly clear that list and add new ones as they’re arrested.’ It is clearly like a machine driven process.

Adam: So there’s a quota element.

George Joseph: It seems like a benchmark, something like that.

Adam: Yeah, they wouldn’t call it a quota, god forbid, but there’s they, they, they give you a hit list and heavily suggest that you clear that list and the only way to clear that list is if they’re arrested, presumably.

George Joseph: Right, and officers are supposed to ask themselves, as we’ve seen these documents that we’ve published, how many arrests have we made this week?

Adam: Right.

George Joseph: Its, its all but saying it basically.

Adam: Let’s back up from LA here. I know that Palantir has engagements with other police departments, New Orleans, I think Chicago as well. The New York Daily News now that it’s changed owners is actually not quite as bad as it used to be. The Daily News had a big profile on someone who has been on the NYPD’s gang database for years. Can you, can you talk about how this dovetails with and is similar to these so called “gang databases” that we’ve talked about on this show that again are another form of kind of pre-crime under the guise of “predictive policing?”

George Joseph: Yeah. So what the gang databases do is kind of create yet another data point that counts against someone. So how it starts is that an individual who, for example, is seen to be associating with other quote unquote “gang members” or wearing certain colors, hot pink-

Adam: Right.

George Joseph:  Apparently in New York for gang members. Um, and being in gang areas, basically so many broad criteria that anyone who lives in a public housing project in a certain city like New York would be considered a gang member.

Adam: Right.

George Joseph:  Even though that status technically doesn’t mean you have committed or even are suspected of a crime legally, but that classification then puts you on police’s radar and other ways that you are never necessarily informed about. So the example in LA with regards to our story, and I think this can probably be applied more broadly, is that with the Chronic Offender Formula being an unknown gang member also counts points against you and thus encourages police to surveil you, um, and do things like warrant and gun checks on you to try to make arrests on you.

Adam: Yeah. And what, and what people don’t realize is that the criteria for being on a gang list is incredibly thin, includes hanging out at certain bodegas, wearing certain colors. Um, it’s, it’s exceedingly arbitrary. And up to the subjective analysis of whoever’s doing the monitoring.

George Joseph: Right. And the other aspect of it is, especially in New York, a lot of the evidence that is eventually brought to court, which I know you’ve talked about on other programs, are for is social media data in which young black men for the most part are taking pictures with their neighbors flashing, you know, so called “gang signs.” Um, now these just maybe things like, like a handshake with a friend that people growing up together have made up to kind of identify their block. But for intelligence analysts seeing that within the NYPD that can then be viewed differently as an organized, coherent gang symbol, which these 14 year olds don’t necessarily know as being sort of understood that way.

Adam: Yeah. I think the average person listening or let’s say these sort of skeptic listening who thinks we’re all just a bunch of bleeding heart liberals, um, I don’t think they quite understand how thin a lot of this criteria really is. Obviously the gang database we talked about is, is very arbitrary. How you get on this bulletin is based on a feedback loop, which has a lot of racist inputs. There are some libertarian groups who’ve, who’ve actually followed people around to see how many crimes the average person commits a day, minor infractions. And it’s, and it’s a lot. It’s quite a bit, I mean, from traffic violations to jaywalking to drug use. A lot of people use off prescription drugs, um, or off label drugs that there is a sense that, ‘Oh, I’m not breaking the law so I don’t care.’ But what people don’t understand is that like if you follow anyone around long enough and you monitor anyone close enough, everybody breaks the law and that the, the, the, the whole problem with this logic, the whole problem with the feedback loop is you come up with a sort of infinitely precise monitoring system that of course people are going to break some infraction somewhere.

George Joseph: But I think the issue though is that when you bring that kind of policing and that kind of high level of scrutiny to a certain population, um, let’s say in most American cities poor black people, um, young black people, what, what ends up happening is that all those smaller infractions that you just get caught for on a regular basis, jaywalking, biking on the sidewalk, smoking outside, things that young white people in different neighborhoods would never be arrested for or cited for-

Adam: No.

George Joseph: Those kind of lead up to, um, days in court that you may miss because you have work or family obligations, fines that you may not be able to pay, lots of small things that then actually lead to a real record which then makes it hard for you to get the job. And so by the time you are older and are sort of being looked at by police as a potential high-level offender, you’ve already established this rap sheet and because of that, you can’t go to the job.

Adam: Right.

George Joseph: I mean it’s already hard enough for black people because of employment discrimination in this country to get a job even when they’re as, or more qualified than similar white candidates. So what do you do? You have to do work to get a living and thus carry around a gun, probably a gun that you are unable to legally hold, but you have to do it to be able to do things like work in the drug world, work in like security worlds, stuff like that. Um, and so we should be honest that it’s, it’s not just police, uh, getting people for jaywalking or something like that. These are people who are sometimes engaging in underground economies with guns with, with things that are technically not legal because they don’t really have any other options. But when police bring this, uh, this kind of surveillance and scrutiny onto them their entire lives, it’s almost inevitable that they would have to go down that path and then they get screwed over even more so for having ended up there.

Adam: Yeah. I definitely don’t want to suggest that they were building these databases off a bunch of innocent jaywalkers. I guess my point is that when you get people on the grid young at 13, 14, you do be, it does begin the cycle that you just described, right?

George Joseph: Right.

Adam: That these sort of minor, low level infractions begin to paint one’s entire life at a fairly early age and one of the major logic behind Broken Windows and even stop-and-frisk and if you really dig into the literature, they’ll sort of tell you this, is to just get people early on the grid. If you look at that data, it really does begin to become a feedback loop.

George Joseph: Right? I think the other interesting thing about, uh, these risk formulas that police are using is that they completely toss out the idea that after you’ve spent time in a criminal justice facility, a prison, a jail, whatever, that you have done your time. So if you’ve been investigated, sentenced, convicted and then spent time in prison, supposedly a way to claim, ‘Okay, now you have a fresh start.’ But what these formulas do is say if they’re a parolee, if they’re on probation, if they have a criminal record, then those are literally points against them that should then be used to justify surveilling them. So the formulas themselves completely discounts the idea that we’re supposed to give people a fresh chance for quote unquote “doing their time.”

Adam: Yeah. You have this pre-crime and post-crime element to it. I want to pivot here to talk about what steps activists and community groups are doing to push back against these forces in Los Angeles and any other cities that would be good as well.

George Joseph: What we are seeing in a lot of cases around the country is that people are just now years after the establishment of these gang databases and these sort of new pre-crime gang sort of practices they’re only now realizing about them and that’s coming through investigative journalism that’s coming through advocates filing records requests and that’s coming just from people piecing together bits of information that they are getting on the street or at court when they’re being hit with these ridiculous conspiracy charges that tie all of them and everyone they know together for one murder that took place several years ago. So in New York, um, what I’ve seen is that a lot of community groups are trying to let people across the city who are in similar situations because they live in housing because they’re poor because they’re not white, to know that these tactics are going on and that they are likely going to affect their communities. And uh, what’s interesting about these operations is that oftentimes they’ll take years from beginning of surveillance to ending in some kind of gang raid or police operation. So you’ll have analysts monitoring these kids as young as 13, 14 years old and monitoring their social media, connecting how they say something about getting revenge on someone or how they make a hip hop video to evidence of gang activity. And so what activists are doing, I’ve noticed, is not only figuring out information about these tactics, but also trying to do teach-ins and warnings to people about how you need to understand how you’re publicly facing social media data as well as other things like your phone calls from Rikers, your letters, that kind of thing, could be used against you at some point a way down the line. And it’s hard because a lot of these kids are really young and when you’re young you write dumb stuff on Facebook.

Adam: Yeah.

George Joseph: And so these kids have the, you know, the penalty of not being able to have exactly full childhoods in the way that other people in more privileged circumstances growing up, you know, wouldn’t be penalized for making mistakes or saying brash or dump things online. Um, so letting people know about that and then kind of organizing to bring light on these practices are, are, are most of what I’ve been seeing activists in cities like New York, Chicago and LA do.

Adam: I guess I’m curious to what extent maybe trying to help out people from not being surveilled as good intentioned as it may be and as necessary as it may be, contributes to this chilling effect where we’re basically having people self monitor, self police. Um, in a way it’s sort of similar to the like, you know, videos telling people like African Americans how not to get arrested by police. It’s like you definitely understand why they, why they exist, but in a weird way, it’s kind of managing the symptoms as opposed to dealing with the disease.

George Joseph: One way that activists seem to be trying to, uh, attack that is by changing the narrative about gang policing and gang raids so that law makers and law enforcement officials aren’t awarded and sort of praised by local media and the public for carrying out these sort of big spectacle, um, operations. Preet Bharara, who is the former southern district of New York Federal Prosecutor, carried out one of the biggest quote unquote “gang raids” in the Bronx a few years ago. Um, he hasn’t really been challenged about that operation, which is much more controversial among the people it targeted then among the larger public. Um, so I think the degree to which these kinds of tactics shift in terms of their public perception could sort of shift how willing law enforcement is to go after just broad swaths of young black people.

Adam: Yeah. I, I know that you interviewed Josmar Trujillo, who is an activist in New York and he had mentioned to Citations Needed, another podcast I have, that predictive policing and these high profile gang raids were basically just a updated version of stop-and-frisk and Broken Windows that they’re kind of a techno liberal friendly version, but they’re, but the sort of effect is the same, which is, which is to create a harassment regime and the media kind of buys into these narratives. It doesn’t ask a lot of questions about who, what is a gang member, who are these people being rounded up? Why are they running teenagers on RICO charges, which was set up to stop Al Capone?

George Joseph: Right. And, uh, the whole ideology behind these two theories, Broken Windows Theory and whatever you want to call it, big data policing or intelligence driven prosecutions is the same. It’s the idea that a few identified hot spots need to be targeted or fixed or cleaned up in order to drive a long-term sort of change in crime trends. Um, whether that’s ever actually worked is obviously much debated in the criminological community. Um, but the point is that both of them understand that certain networks and communities face far more difficulties in terms of employment, in terms of just getting by. And so the response is to set up long-term surveillance practices on them that guarantee they’ll be prosecuted and locked up. That’s the kind of way that major liberal cities are dealing with their public housing population at this point.

Adam: Yeah. Its, its, in many ways criminology is limited by its own kind of epistemological scope. It says, ‘Oh, we’re just dealing with the problem as it is,’ and there’s no broad questions as to whether or not the very nature of it is racist. Yeah. It reminds me of a, of a thought experiment a professor of mine in college used to explain the sort of institutional versus interpersonal ways we interpret these things. Uh, which is that if there was, for whatever reason, if there was a psychiatrist in the Antebellum South and the 1850s who was, who had a patient who would whip slaves with a whip and he was feeling really bad about it and filling neurotic about it. I could sort of manage him, give him prescription drugs and talk him through it but I could never really questioned the institution of slavery. There’s this sort of institutional racism to the very nature in which we deal with crime and I think people who look at this like it’s just, oh, it’s just this cold computer, its just this algorithm, that they don’t understand that like by the very nature of trying to target certain neighborhoods and trying to try to jack up your numbers with quotas like you’re advancing a system which is manifestly racist however you want to cut it.

George Joseph: And one thing that, um, people who aren’t activists you often hear of, but just organizers within their communities, especially in public housing in New York will say, is that, ‘Look, you’ve taken away our playgrounds. You’ve stopped our kids from having like basic places they can hang out and engage in, in a healthy social way. Um, and we’ve been asking you to help us for years. You’ve been gentrifying neighborhoods and not giving us anything to get by.’ And then years later you just arrest all of them, um, without ever having ever tried to help on the social side of things. So I mean that’s kind of just the choice that authorities in the city are making. Um, and how, how they’re able to make that choice I think reflects on how much the general public, uh, seems to be willing to, uh, tackle this issue.

Adam: Yeah. And this was seen the most starkly of course in Chicago with a $100 million police training academy, while Rahm Emanuel has closed, the Mayor of Chicago has closed dozens of schools that there’s, there seems to always be money for policing and never money for anything else. All right, well I think on that note, George Joseph, thank you so much for coming on.

George Joseph: Hey thanks for having me Adam.

Adam: Thanks to our guest George Joseph. George is a writer at This has been The Appeal Podcast. Remember, you can follow us on Twitter @TheAppealPod and on Facebook at The Appeal Podcast. I’m your host Adam Johnson. This show is produced by Florence Barrau-Adams, executive producer Sarah Leonard. Thank you so much for joining us, we’ll see you next week.