 Hi, I'm Scott Jones. I'm acting director of electronic frontiers, Georgia and So tonight's presentation is the trouble with gunshot detection systems We want to talk we want to kind of back up and talk about what our gunshot detection systems are And is there really anything to be concerned about? Is there any really real problem? I mean what we found out we do have a gunshot detection system by the name of shot spotter That's been installed in Atlanta and we found out that one is being installed in Macon, Georgia And we also found out that that's being That's being funded by covered relief money So the covered relief money has accelerated the installation of these systems So if there is a is it if there is an issue or a problem we need to understand that And it does add some urgency because suddenly there's all this money flowing from the federal government and Really no no real direction about what it can and can't be spent for so I want to go ahead and ask our speakers to introduce themselves first Let's start with Alex. You can go ahead and turn on your microphone if it's not already on and and get it started Thanks, Scott. Can you hear me? Okay? Yes Hey all my name is Alex. I use she and her pronouns based in Chicago I work as a deputy campaigns director at Action Center on race in the economy. Some people know it's this acre And with that and outside of that I'm also part of Chicago's local Defund campaign and Chicago's campaign to cancel our shot spotter contract And I'll pass it over to Ali Hey everyone, I'm Alejandro. I use they them a pronouns I'm a co-director for Lucy Parsons labs in Chicago And I'm also part of the campaign to cancel the shots for our contract within Chicago. I'll pass it over to Jonathan Yeah, my name is Jonathan Manis. I Use he him pronouns. I am an attorney at an organization called the MacArthur Justice Center in Chicago Okay, Alex if you could go ahead share your screen we can get started with the presentation Okay, so this Presentation is a copy of her presentation that we did back in December called public safety is not profit And we did this as a way to introduce Shot spotter and a campaign against that spotter to a more like national audience of organizers So just so y'all have some context for where these slides are coming from This agenda is the agenda that we use we obviously will not do all of this together We're mostly going to run through The first three things on this agenda. So I'll do very quick on surveillance capitalism, which I'm sure all of y'all are very familiar with Jonathan will give an overview of the research That's been collected and that we've been using a sport our campaign in Chicago And then all I will take us through some of the things about Chicago's campaign And then hopefully we will have time for a question and answer at the end So I want to back up a little bit and talk about how I got introduced to this shot spotter So I mentioned that I work at Acre And the focus of my work is around the relationship between policing and Wall Street which means that I Get to do really cool work that takes a more corporate corporate angle and corporate focus on Policing and through research that I was doing at Acre with this framing in mind I came across shot spotter. This was maybe like a year or two ago now And we started like learning more And thinking about Surveillance tools that police departments use like as corporations and then we know that they are but I think that this gives us like another Another target and another frame and another angle to lobby on Saying that Shot spotter costs our city in Chicago nine million dollars hits But when you say that Chicago is paying shot spotter nine million dollars a year I think that hits a little harder and so I think that one thing that our campaign has tried to do is Yeah, just like reinstall and like reinforce the fact that these tools that police departments are using are For the purpose of like their own of like corporate profit not for public safety Because things that like campaigns like ours are fighting for Around defund or around abolition those things are not profitable, but reform and Predictive policing and more effective policing those things are profitable for somebody and so We think it's really important to name those things and our messaging and the way that we talk about shot spotter And eventually how we talk about the gunshot detection industry as a whole and so, you know We understand surveillance capitalism as a ways in which private human experiences are collected computed and sold off to private businesses As behavior prediction and you know, we I think Jonathan will talk more about how exactly shots water works But that's essentially what we are seeing We're seeing gunshot detection technology do right there's this idea that like by installing a bunch of microphones in our neighborhoods that it's going to like encourage people to behave better and Data that shot spotter collects on The sounds that it hears is is then fed back to the police department that inspires More policing in our neighborhoods and also fed back to shots water and shots water gets to say that they like Counted all of these alerts and this is why you need to sign more contracts with us and so There's just some of like I think some of the more important points that we want to make sure that we drive home We're thinking about shot spotter not just as like a tool that police departments use but this is like a tool and the larger like toolbox of capitalism and Lastly our goal is not a better shot spotter. Our goal is the removal of surveillance technology as a whole and With that I'm going to pass it over to Jonathan Everyone It's nice to be here in the space with you able to hear and speak hopefully be heard so So I'm going to present a little bit about You know what shot spotter is the problems that we've discovered through You know researching what's been published and some original research that we've done the coalition in Chicago and at MJC and Talk about some of the costs and risks of this form of police surveillance so Yeah, if you can take ahead the next slide So here's how here's how it works some folks may already know this but I think it's useful just to kind of level set So the way so shot spotters What it claims to do is it claims to identify the sound of gunfire in an area that's covered by its system and Determined the location of the gunshots and send that information directly out to police very quickly Within a minute of the loud noise That's it. That's its claim and the way this system works is that The city that contracts the shot spotter Tell shots butter what area they want to Surveil part of the city and the shot spotter sends out its engineers or technicians and they install microphones About 15 to 20 per square mile. So whenever a few blocks Sort of blank at the area I Just want to I need to go check on three jokes real quick. So I need to mute my mic and then reshare if that is okay with y'all Alright, thank you. Sorry. I'll pick it back up. So so the way it works is that these You can see on the on the left here. There's a picture of what these devices look like there's there's two models that we're aware of this they look like and The way it works is that these microphones are always listening always recording each device actually stores 30 hours of audio locally on the sensor And has some software on the sensor so that whenever there is any kind of loud Impulsive sound any shots by likes to talk about bangs booms and pops anything loud it Sends that audio clip back to shot spotters computing system by the cell phone network with some metadata and Those audio clips are then processed through two algorithms one which Attempts to triangulate the location of the sound by looking at the timestamps How long it took the sound to reach each microphone and another one that crucially? Proports to classify the sound as gunfire possible gunfire something else these algorithms Are don't don't work all that well at some point over the last ten years shot spotter Started hiring sort of call center style staff who review each audio clip and some visual representations of the Alert and decide whether to trigger an alert and that person that's like sort of call center staffer job Qualifications are you know basically high school diploma and some call center experience They're the ones who are actually triggering the Alert and effectively dispatching police to go out Chasing down supposed gunfire. So if you can go to the next slide. So here's the here's the trouble So a system like this it's been on the market more than 20 years You figure like they would have tested and validated it And there's really two kinds of testing validation You need one like basic sort of fundamental validation in general does the system do what it claims distinguish gunfire from other sounds? located accurately and then in a particular implementation Is it doing that and are like these individual call center operators like reliably distinguishing gunfire from other sounds? The crazy thing is that shot spotter has never done that kind of testing at least not that we know of publicly There's never been any published study testing the system against Things like fireworks cars backfiring Construction noises tires blowing out things that like you and me might confuse for the sound of gunshots So we we there's really no Scientific empirical basis to judge how easily the system is fooled and There's been no independent audit of shot spotter system apparently despite requests, so we just there's an NBC news article recently that reported that sort of Surveillance trade publication called IPVM They're kind of I think they style themselves like the consumer reports of like surveillance Tech and they asked shot spotter. Can we do some testing of this and shot spotter said no, so that's interesting and A couple other key points. So these human operators call center staff They play obviously play a crucial role because they're the ones who are getting that output of the algorithm deciding whether to trigger the alert they they're They apparently Are supposed to follow an internal classification protocol known as the quote classification continuum Shot spotter refuses to disclose that it calls it a trade secret So we can't even know the protocol they use to classify things Which is pretty that's not how science works. I mean imagine like a DNA lab that didn't tell you how they like Did their science that wouldn't that wouldn't That wouldn't be legit so They keep that secret of that there were there was one expert who independent expert who was able to review that subject and on disclosure agreement and All he was able to say publicly was like anybody who's using the system should be able to see this document because you realize how subjective the system is Another interesting note just for folks who are You know nerds about technology and algorithms So this is the way the algorithm works is that it actually transforms the audio into like a visual depiction And then they use like a visual machine learning algorithm to try to determine Classify whether that visual depiction looks like a gunshot or something else But the trouble is that the whole algorithm was trained on just subjective determinations about whether an audio clip is a gunshot or not It's not actually trained on known gunshots known fireworks So the algorithm is trained just to replicate whatever errors, you know humans listening to Sounds and making judgment calls would make so That's that's not how you design a proper ML system, even if like it would It would otherwise work so next slide please Okay, yeah, so so so that's like the general sort of fundamental I believe right reliability most cities that deploy shot spotter don't actually test the system once installed So this is an example in Chicago Chicago didn't allow any deployment qualification testing before the server one service one lives that what that means that they didn't even allow They didn't even do like test firing blanks to see whether shot spotter picks up actual gunshots or guns being fired off and Is is triggered by those so you know let alone testing for false alerts to things like fireworks and cars backfire Next slide so So that's like a little bit of backgrounds like what what do we find out? just by looking at police data so You know without actual like control testing we can't know shot spotter can't know nobody can know what like the True rate of false positive alerts are but what we can know is what do police find when they show up at the location shot spotter Sends them so shot spotter setting police out a specific location it's doing so quickly and then police are going out in response and What we found in Chicago was that? About 90% of the time police go out and find no evidence of any gun incident at all So no shell casings nobody armed no shooting victim no evidence of a shooting nothing to corroborate gunfire and This happens a lot so in Chicago. There are more than 25,000 shots bought alerts a year That works out to more than 60 police deployments every day Where they go racing out in response to shots bought alert and don't find anything to corroborate a gun incident They might find a person there who they stop and frisk We'll get to in a second, but they don't find anything to corroborate a gun incident and similar numbers have You know since have been Found in other in other cities now So, you know just to give you a few examples in Dayton a journalist there did Similar kind of study found that of all the shots bought alerts Only about 5% led police to Report finding evidence of any kind of crime when they arrived at the location less than 2% of the alerts led to an arrest Similar story in Houston less than 1% of alerts led to an arrest Less than 20% of shots bought of deployments led police to record any kind of offense not just got a fence with any kind of offense when they arrived in Little Rock, Arkansas 2026 alerts led to only eight arrests and Seven suspects identified Just over 10% of those alerts led police to record any kind of fence So, you know consistent across the country the vast majority of Shots bought alerts are leading police to find nothing police should expect when they go out on shots bought or Deployments that they're not going to find anything Most of the time Next slide so You may have heard the claim From shot spotter that they're 97% accurate. This is a marketing claim that they continue to repeat they actually commissioned a Consultancy to sort of to quote audit this number That that's not an accuracy number. This is I think a deceptive and false marketing claim So this staff first of all, it's not based on actual testing of the system, you know, they're when they claim accuracy It's not based in any actual empirical testing the way they get to this number is Let's play it this way. So they start by assuming just assuming without evidence that every alert is Corresponds to an actual gunshot. So they start by assuming 100% accuracy and then they only count something as an Error if the police happened to send them a complaint about a particular Alert that the police think they got wrong. So for example if the police say hey We had a shooting and you missed it and then shot spotter goes back and says, oh, you're right. We did we missed it that counts as an error and And and and then this subtract that from the hundred percent. So this is just a tally of customer complaints This isn't an accuracy rate Unless we all get to a start from the assumption of a hundred percent accuracy, which isn't how science works. So So this is a misleading stat what the data show is that most of the time police go out and find nothing and You know whether in that 90% of alerts that The question is like what's going on that 90% of dead-end alerts and You know, it might be that some of those the police, you know There was a gunshot and you know, the police show up and just don't find any evidence of it You did despite looking for shell casings or whatever it seems Very likely that a large proportion of those are just maybe they probably weren't gunshots at all But you know again without testing you can't know what we do know is that most of the time police will find nothing Next next slide. So, you know consequences like the Important to note that like all of these like dead-end deployments aren't cost free So this these are situations where the police are being effectively told by shot spotter like that you're going to location where shots were Just fired be on high alert for somebody who's armed and just fired their weapon So this is creating sort of High-intensity situations where police are on alert What the inspector general in Chicago found they did a report after we publish ours they found thousands of Shots butter alerts that lead that led to a stop-and-frisk an investigatory stop pat-down questioning search for the person That's almost certainly an undercount because Chicago keeps terrible records. So So oftentimes the police won't Link the investigatory stop to the shot spot or alert in any way that you can get from the data Sometimes they don't I know from first-hand accounts. Sometimes they don't even write up the report of the best for a stop So This is almost certainly an undercount Also, what police in Chicago started doing is when they have to write up their reasons for doing a stop-and-frisk they're starting to say things like This person we found this person in an area that has previously historically had a large number of shot spotter activations So they're in a high shot spotter area Not that they're actually responding to a shot spot alert But just like we have the general impression. There's been lots of shot spot alerts in this area so shot spotters like functioning as a sort of tech wash justification for stopping for us just because it's installed in an area and And the inspector general found that It appears to be changing police behavior in that way Next slide I'm sorry. Can you can you go move ahead to the next other way? Awesome. So So so so thinking about how it's deployed again, this is you know Chicago experience But I think this holds true in lots of cities This on the right is a map of where all the shot spotter dispatches were That that we used in our in our study and if you go to the next slide It will show you the racial demographics of Chicago. So basically the south and the west side you see are where the Highest proportion of black and Latinx people live in Chicago. That's where a shot spotter is deployed Basically only in those parts of town It it's if you if you rank basically Chicago police districts by the proportion of black and Latinx people It's the exactly the 12 districts that have the highest proportion of black and Latinx residents that are covered by shot spotter And the lowest proportion of our residents fully 80% of black Chicagoans are Living under a shot spotter surveillance footprint only 30% of white Chicago's are so it's pretty stark Um Next slide This is this is sort of just another way to visualize the police districts. So On the left and blue these are the police the demographics of the police districts that are covered with shot spotter The red bars are the proportion of black and Latinx people in the district and you can see how the City chose to deploy the system Okay next slide so Shot spotter has you know, it promises to Do a couple things Actually shot spotters marketing and they sort of had different justifications for like what is the system for? But I think the main point is to reduce gun violence. I think that's why People are looking to the system and there's very little evidence if any that it does that so a study from johns hopkins university looked at 68 counties that had installed shot spotter somewhere in the county over 17 years Looked before after shot spotters installed found no difference in homicides or arrests um a study from st. Louis uh did a similar Comparison looking at areas with shot spotter and demographically similar areas of shot spotter Found no reduction of violent crimes What it did find is like a massive increase in the number of police deployments hunting chasing down for Chasing down gun shots because shot spotter just generates so many alerts, but that didn't generate any reduction in Violent crimes or any increase in arrests if I recall correctly um Shot spotter has made other kinds of claims for what its system is for sometimes they talk about You know, this is a way to get police out into the community and you know improve relations with the community That feels like a very strange justification that you you want cops going chasing down like Lazing into a neighborhood um Responding to supposed gunfire as like a community policing strategy that doesn't really add up More recently shot spotters marketing has focused a lot on the idea that actually what shot spotter is about Is about saving lives by getting police out to shooting victims faster by pinpointing the location of shooting victims And getting police out faster so they can render aid um You know a few things to say about that one is that a tiny fraction of shot spot alerts are actual shootings um But beyond that there have been two studies trying to look at this one study from a hospital in camden another Recent study october 2021 From a hospital in harford looking to see if there was any difference in patient outcomes shooting victim outcomes If the person was shot In a shot spotter area versus a non shot spotter area or if the police were alerted by shot spotter to the Shooting or if they weren't and they found no difference. Um, this most recent harford study found No reduction in pre-hospital times or transport times um Found quote no benefit to having shot spotter. So this again is like I believe shot spotter trading on like an intuitive sense of how this like would work in like a perfect world But in practice, it doesn't seem to help It's it's a good pitch, but it doesn't seem to be true Next next slide So just to sort of talk about the cost. Um, so the way so there's the dollars and cents in Chicago, it's nine million dollars a year. It's about 90 000 dollars per square mile. Um I think you have to add on to that the cost of all the officer time Responding to shot spotter alerts. So shot spotter isn't a system that like Makes policing more efficient. It actually increases the number of police deployments without any Uh anything to show for it. So it's actually, you know, increasing the Demand for additional police officers Um It leads to stop and frisk. It creates these high intensity volatile interactions with police Um, it also creates a risk of rockfall convictions. So in Chicago, um, a man named Michael Williams Was arrested on charges of murder and held in jail for 11 months Largely on the strength of Shot spotter evidence He was accused of uh, uh, I guess had a murder and 11 months into his trial Are his criminal proceedings The prosecutors Um conceded that the shot spotter evidence couldn't be used against him wasn't reliable enough can be used And immediately dropped the charges. So he lost. Uh, he's an innocent man um, he didn't commit the murder and He was robbed of almost a year of his life Contracted cobit. He's a 65 year old man with chronic obstructive pulmonary disease um Suffered a great deal Because of shot spotter. Uh, there's another man named Sylvan Simmons who was prosecuted In rochester, um after he was shot by police officers on the theory that he had shot first The only evidence that he had shot first as far as I understand was um Shot spotter supposedly detecting an additional gunshot um That additional gunshot was not detected in the initial shot spotter report Um, it was only detected after shot spotter went back and apparently reanalyzed the audio twice um It never found any bullet or shell casing or other evidence that he shot the police first um, he's now suing The rochester police and shot spotter Or putting him through that ordeal So, uh, so there's risks here about how the shot spotter until we used in trials. Um, a couple other concerns. Um Shot spotter, I mean the whole design of the system is about A way to send police out without relying on people calling 911 and actually one One finding from the st. Louis study and I think another one too is that people actually end up calling the police less one shot spotter is um After shot spotters rolled out in the neighborhood. It's kind of not clear why but in any case it's um It seems like the system is sort of cutting the community out of public safety processes um And it's really training officers to follow the tech around and trust the tech rather than you know people who are willing to call the police, um So and then one one um One last Thought is is the concern about sort of false metrics and inflated gunfire stats. So um Shot spotter generates all of these alerts and that feeds back into uh other Processes that police departments use so in chicago They have like a comp stat system where they are held accountable to meet certain metrics one of which it appears is Um the number of gunshots in an area So shot spot So this to the extent that shot spotter is actually just alerting to things that aren't gunfire. You have police chasing phantom metrics um in chicago also The the shot spotter gunshot stats feed into the predictive policing system Um, which shot spotter actually owns used to be known as hunch lab I think they call it shot spotter missions or something like that now. And so basically shot spotter is um generating predictions forecasts about where crime is going to occur based on these gunshot action Alerts and leading the police to deploy More resort more officers into areas based on the Fault that you likely fault the results of the shot spotter system. So It can basically skew The way that policing works And creates sort of a tech tech wash justification for More More policing and Um Forms of policing that I think a lot of folks have been organizing against around the country So, um, I'll stop there. Sorry for taking up so much time Let me pass it over to Thanks, Jonathan. Can you hear me? That uh Sweet awesome. I'm gonna talk a bit about what we've been doing in chicago with us far Um, some of these slides might be a little out of date So I'll try to point out some things that we've done after these slides are created But initially we had three core campaign demands within the city of chicago So we wanted the city of chicago and the chicago police department to cancel their contract of shot spotter This contract that the city has at the moment can be canceled at any point in time Uh, and we we use that a lot in our messaging because initially what we wanted to do was we wanted to get it canceled in the middle of last year and then we found out that the city had already renewed the contract without any Any public comment any sort of public inquiry whatsoever the year before and uh So that's that's really why one of our first uh goals here one of our first demands is getting this contract canceled Um apart from that we also wanted to have uh an act of divestment happen where the money that was used Initially to pay for this system across various years and up until now is redirected to community Constructed ordinances. We have something called the peace book here in chicago that was created by a lot of youth here and We initially wanted the resources to get redirected to initiatives and efforts like that in order to effectively address gun violence In a way that's that's community based and community centered We are explicitly abolitionists. So we also didn't want any sort of police involvement whatsoever Um, and then the third thing is conducting an independent audit of shot spotter and its impact on communities of color As jonathan mentioned before there hasn't been any publicly disseminated peer reviewed study on the efficacy of shot spotter these questions of You know accuracy and efficacy Are have been really difficult to penetrate into because the company in and of itself doesn't like showcasing, um, it's its own false net negatives and false positive rates Really any sort of uh analytical tool or device that could make this Easier for us to understand what's going on has been denied and I think um, the other uh point here too is that That We we've been able to see research a lot of research growing that shows it doesn't work Um, so we'd like to just see this independent audit Come through and see what sort of harms have already been created in chicago And might need recompense in the future Next slide Well, and so this says current focus again, uh, this is current focus back a while ago So it might have shifted a bit. Um, initially we wanted to get 26 city council members to support canceling the contract shot spotter Um, at the moment, we're still working on getting more, uh, camp of more city council members involved Um, I forget exactly the number that have uh more or less support to this I don't know if jonathan or alex you remember the number But if you do pop it in the chat, uh, and we are also hoping to ensure that the funds from the contract are reinvested In these sorts of oh, thanks ed. We've got 11 at the moment. Um are reinvested into these community driven efforts Our strategy. It's been multi pronged. We've taken different sort of actions Initially, we wanted to do more base building through canvassing and political education and target awards, which is something we still do I mean, um at the moment, uh, this this presentation in and of itself sort of led into some other More focused presentations for various community groups that have been happening We want to activate the base this sort of base the pressure and move the target alterment to support our campaign But also even just beyond that to have a to spread this political education Across communities that are most impacted in chicago So they can gain this this sort of knowledge and use their own autonomy in deciding what sort of actions need to happen um, and finally it's just really shifting shifting the narrative away from The way public safety is talked about now More or less. We've noted that the notion of public safety is this sort of abstract you know entity that is usually Restrained by police to mean things like oh, yeah public safety is Resting criminals and bad guys or whatever sort of a you know notion they might be using Um, whereas we really believe that public safety is something that needs to be determined within the level of community Um within the level of neighborhoods that are being surveilled and so forth in order for us to World build something a little more interesting and different than this rapidly expanding surveillance process here in chicago Um next slide please And so here's some of the things that we've done. We launched in june of 2021 I believe and so this this happened. Um after adam toledo was murdered by the chicago police department following a shot to butter deployment it was you know, it was really hard to hear that sort of news and I think a lot of us at the time were doing organizing and movement work In different spaces that are related to abolition, but also started Dipping our toes into surveillance and when this happened. This was really just like a big call to action for a lot of us Um in july the city revealed that the contract was renewed in december of last year At the time we thought that the contract was going to be up for renewal But uh, this is what I mentioned before it was a little a little bit, uh Uh, what sneaky what was what happened when it got renewed without any public inquiry Um, and we had a rally in march in the neighborhood of little village in chicago to honor adam toledo is from Um in august the associated press reported that cpd and shot spotter framed mr. Michael williams for murder Um, and jonathan spoke about that a little earlier And we also had a rally and mapping in the angle would neighborhood in chicago Um at their sdsc which stands for a strategic decision support center That's a chicago police department location where data is fed into various predictive policing algorithms and so forth um after that the office of the inspector general released a report that corroborated In cars of justice center findings in september we introduced a resolution to host a hearing on shot spotter Or sorry a resolution was introduced to host a hearing on shot spotter I made that correction really quickly because it sort of Happened without our involvement and mostly with the involvement of Individuals who were in city council and already very pro shot spotter Um, but we had the opportunity to give public comment and to make the conversation much more nuanced than they were trying to set it up to be Um in october we had a people's hearing on shot spotter. Uh, this was done primarily digitally And we were able to talk about shot spotter Get community comments and so forth Alderman briefings on the budget amendment with micarta justice center and oig representatives Um an amendment to cut the contract funding was introduced but wasn't heard by the budget committee Which is one of the uh city council committees here um, and in november of last year Uh, we had the public safety committee hearing and action at the chair's office Which is which is actually uh, where all of this happened. Um, I mistakenly said this happened earlier But this is where uh, we made the conversation a little more complex Um, and is there another slide after this? Cool. And so here are just some of the tactics we've taken. They're roughly broken up into three different categories political education A city council strategy and the sort of materials and tools we've used So in terms of political education, we've done things like we've held rallies in person. We've had press engagements multiple times Um, we've been doing a lot of teachings within various communities here in chicago Because our coalition is composed of I think over 60 endorsing groups at this point. So we've uh, been able to really work with other organizers and Produced materials and resources that are are useful for the people who are represented within various communities Um, we had a people's hearing online. We've again done a lot of canvassing graphics and social media campaigns and so forth Um, not listed here, but I would probably throw it into is we've done a lot of Freedom of Information acts Sent out countless of them to get a lot of public records data that we can use in trying to further understand how these Decisions were made at the city level and the implications of said decisions Um, and in terms of city council strategy, we've done things like emailing the reported findings and data On shot spotter from some of the research that was mentioned today We've had meetings with other coalition members pretty constantly um, I think uh Yeah, alex and and I know there's like the joke that we thought this was going to be like a two to three month campaign And we've been added a little over a year now, and it's still going um, we've targeted emails phones and tweets, uh, that we've strategically Um looked at and we've also done alderm and briefings during the budget season As well as raising public comment during these committee meetings in order to Try to better represent the voices of ourselves and our communities when city council makes decisions and Doesn't put an agenda out early enough And then we also had a series of materials and tools that we used so we have a petition That constantly grows. We have a toolkit. We have calls to action social media graphics wheat pacing firing, you know, the whole shebang We call email tweet templates That we send out every now and then when we need more engagement from the public and calls to action um And most recently too, it's not really described here But alex alluded to this at the beginning. We've started working at the national level now So we work with organizers that represent different regions here in the u.s And have also started seeing shots butter begun to be installed within their communities and Follow these same sort of patterns we've seen in chicago or at least those that are similar enough Um, yeah, and I think that's the last slide. I'm not sure but maybe one will surprise me Cool. Yeah that ignore that. That's not Yeah, except the the call for questions. Yeah, and I see one in the chat right now chic uh from kareem chicago pd is in favor of this tech I don't understand why yeah, yeah, I mean Whether the police themselves are in favor of it is also questionable. I think uh, I don't know if uh, anyone wants to talk about that jonathan or alex You had mute jonathan by the way um Yeah, I mean, it's the the police brass are certainly um Um behind this uh and um And the mayor seems to be um behind this. I think there's there's probably um a question about whether sort of officers On the streets think it's worthwhile. Um, we've Folks in the coalition have done a lot of canvassing and what I heard from another member of coalition. I think ed. He's here is that um Actually in the course of canvassing we've come across uh, maybe four or five police officers Who we just happened to approach and most of them have said yeah, this doesn't work and I think two of them actually signed um Signed on to like they call it the cancel the contract. So um I I think that the um There may be this may be a situation where it's in the political interests of Leadership to be seen to be doing something. Um, and shot spotter has this kind of slick marketing Like who wouldn't want to have a system that like precisely detects gunshots and sends police out quickly Seems like it seems like a good idea if you don't think about it too much. Um, but uh So it's a way to Seem like well, we're doing something but turns out it doesn't seem to actually help So And I think if I can add on to that too, I think we've really seen this uh growing tension Um throughout time, but maybe most recently once uh, the pandemic started in between Dealing with crisis scaling it dealt with fast and trying to see that happening um versus these different sorts of uh, primarily, you know, not often not always but oftentimes tech companies that have a Acclaim to a solution. Uh, that could just be bought at an easy price. So I I think there is that like expectation for city council to do something and And you know, these are a lot of the times these are social questions that require A deep understanding of the history and systemic issues in at least in chicago Um, rather than do that work. It seems like a lot of people just want to buy software unfortunately Um, and kareem also had a question about what are the political roadblocks that have gotten in the way of the campaign um Yeah, I mean, I think uh to To jonathan's point a little earlier one of the issues that I've definitely noted is how often Uh city council and shot spotter themselves move the goalpost So they they tend to you know, say one thing like this is to detect gunshots and then We will argue. Well, you know, the data isn't saying that and then a few weeks later We'll get something Well, you know, we'll start noticing that the argument has shifted into being like We we do this because it gets officers faster to a location where someone could have gotten shot And when you push back against that enough, you know, those Those bits of rhetoric tend to crack too. Um, when that was going on, I think a week later There was a city council meeting where the superintendent, uh, superintendent david brown of the chicago police department Actually said they didn't have any data showing, um, how fast a deployment got to its its site And so like, you know, it's uh, it's questionable when you when you hear these arguments and then realize that the city was actually never looking at it Um, I don't know if that totally feels your question too and I'll open it up to jonathan and alex as well Yeah, I I think, you know, they'll they'll try to find anecdotes and say, oh, there was this like one case where maybe it worked It won't give you the full picture so, um Yeah, I think I think politically the political roadblocks is that um folks elected people and city council the mayor the police chief they're all facing this kind of political climate where people are really concerned about gun violence, um and perception of increase increasing violence, um and Uh It's it's that that increases That makes it um increases like the burden on folks like us to persuade them like actually like if you really want to just gun violence like Don't spend your money on this spend your money on something else um and uh And I think it's an easy It's much much easier politically for Folks to say like well, we'll just like spend money on something and you know It feels like we're doing something. I mean, there's one sort of remarkable example in um Syracuse new york which had shot spotter. They were facing a really serious budget crunch and they decided You know the shot spotter system is not worth the money. So they um stop paying cancel the contract and then somebody challenging the mayor For reelection. It was like how dare you cancel a shot spotter There was like a shooting last month and you've turned off shot spotter. You're like disarming the police um And then the mayor like Immediately has a press conference with the police chief and it's like we're going to be turning shot spotter back on and we're looking to expand it and Nowhere in this conversation is any discussion of whether shot spotter has actually been effective in addressing gun violence And you know it it seems like that this may be a dynamic that's occurring in atlanta right now so atlanta did a pilot in 2019. I believe with shot spotter and concluded that It wasn't worth it. I think that they did a I can drop into the channel find it but they did a report after the pilot And they found that over six months Shot spotter was linked to I think five arrests Worked out to like $56,000 per arrest and they decided that it was not worth the money And now for some reason three years later It seems like they're considering Buying back into the The system so for what else but this is also like the financial model ideas like it's it's meant to be sticky So once you sign up there's a free trial you sign up whatever Um, it's hard to it's hard to stop paying. It's like a subscription It's like netflix or something you sign up and And just kind of keep it I wanted to jump in and say to to my knowledge it actually is active on the southwest side of town or in the west side of town Um, so the problem is that we had a private pay or Georgia power come in and and pay but of course they own they own some of the uh the utility polls Um, but they paid for that. They've also paid for Some of the um the the license plate readers the Alpers Um, so it's harder to get accountability when you have a private entity paying for it Yeah, definitely I I think it I Ali might know more about this than me, but I think in Chicago um maybe shot spotter and other surveillance was originally paid for Through funds that didn't go through the ordinary appropriations process through like civil asset forfeiture funds. So, you know Again, that is revealed without any public Input or budget process Yeah, exactly So a lot of the a portion of the funds a substantial portion I should say come from civil asset forfeiture money here in Chicago following the war on drugs And so that wasn't reported in the uh in the police budget When these purchases were made uh, in fact a lot of these purchases weren't uh reported in the police budget ever because they were managed by this uh This this sneaky little other group that no one in our committee had heard of and a few older people seem to know you had heard of um That we uh got through for you um Yeah, sorry. I'm just running through the chat too a lot of a lot of good activity here So even in our own group electronic frontiers georgia, there's a debate going on about Whether whether we should try to get rid of the system or should we try to work with the company to make it better? um and uh So we don't have really even complete agreement within our own group So I wonder if you can speak to you know, if we if we try to make this work with them to make the system better Is that is that even ridiculous to suggest or you know, I think at some point there's going to be more regulation around how um How police surveillance is done and I can't think of too many regulations that wouldn't cost a vendor money So do you think they would put up a would you think they would accept willingly accept that? Or do you think they would put up a fight? Uh, if we try to to um put some regulation around it Take a stab at at least a little bit. Oh, no, alex go ahead, please Um, just just some of the first things. I like uh, just on the first question of like Do we make it better or do we get rid of it? I think one of the issues One of the many issues with shot spotter is um, something to what jonathan was saying like shot spotter doesn't reduce gun violence So it's like even if it detected gunshots accurately, it's not preventing gunshots So we're just again like putting more money into the system of policing um rather than putting money into like Things that would serve as a social safety net and like actually reduce policing and I mean communities want like Communities have named the things that they want in public safety. Yes. Sometimes they say police um, but they also are saying we want housing quality schools like they want healthcare they want like mental health care and I think that um Yeah, so, I mean, I think that is like my thought there, uh, and I I'm Uh on the regulation piece. I think I still I'm like if we can get rid of it, we should get rid of it. Um, I I think like we saw in New York Uh with the post act like their form of regulation came and just in In terms of transparency and so all nypd had to do was say like These are the tools that we use and this is what they're supposed to do and that was like Regulation, um, but I would also like love to hear from Holly and Jonathan Yeah, I mean, I fully agree. I I don't think that this is something that can be reformed into being functional or good for people. Um, and so I think, uh There's a there's a lot of layers there, right? Like I think that if you only focus on the tool as like Some sort of an object or an artifact that exists outside of the reality that we live in That like maybe there's a case there for making it better, but that's just not how it functions, right? Like this is a tool that's always been an extension an extension of the system of logic of incarceration It is largely as Jonathan showed deployed in these, um, locations that are already primarily black brown low income Uh and creates in and of itself creates more policing data that ends up um I guess incentivizing other policing initiatives to constantly send more and more Uh police more and more deployments and so forth to the same locations Um and ends up sticking itself into sort of this circular logic of constant incarceration and I think, um You know, there there has to be a big sort of a Ethical and moral question that you think of when you think of technology is like this. Um, what is better, right? Like what is good is uh I I grew up in a neighborhood that had a lot of gun violence It was something that made me scared as a kid. So I take this issue quite seriously And I just don't see this sort of a tech being functionally useful because it doesn't it doesn't One it's um, it's retroactive, right? Like That's already been created and I think in this coalition. We're interested in larger Solutions that are preventative rather than retroactive solutions that can functionally, I guess uh And identify the sort of a more systemic uh issues that are producing things like Um gun violence and so forth which is largely, you know, criminogenic from things like poverty or or an increased police presence in and of itself has been shown to be criminogenic as well. And I think um Yeah, I think just at the at the end of the day it would be sort of a a nice sense of um Confidence that might be placed on believing that a system like this could be better without noticing all the political structures That have placed them in neighborhoods like mine to begin with, right? Like I think at the end of the day there were incentive incentives to do this And while the narratives might be You know, we did this because we want to increase public safety. I can't help as a as a brown person I can't help being like we got to control these fucking brown people, right? Let's put more tech in their neighborhoods That's what it sounds like to me. And that's I think why I have an issue With uh, these sort of tech based solutions where the solution is just let's put some sort of monitoring system and arrest them When they got out of line and not put it anywhere else where there's a wealth or whiteness I don't think that that's something I would feel comfortable supporting So I did want to mention I mentioned it up front when we started but I want to mention it again uh, we just found out in the last two weeks or so that uh That there isn't a system also going on in Macon, Georgia and Macon, Georgia's to the southeast about 90 minutes southeast of Atlanta Um, but it is it is considered a separate. I guess a separate market in terms of You know television and and and things like that Uh, so it is it is Close and yet far away in a sense But we you know, we also found out due to reporting from a local tv station that the funding is coming from The uh, the covid relief money. It's not the infrastructure bill. It's the american recovery act And so what happens is all this money is raining down um On the smaller towns and there's there's no real restriction about how they can spend it And you would want to see money like that being spent on officers uh, you know squad cars um uniforms radios just traditional stuff and and not not necessarily on toys that that Have, you know questionable utilization, um Well, you could say that about any surveillance tech. I mean the problem is that it's Uh, and the vendors the vendors love all the the free money rating from sky But there's no controls out there So I think that the problem is quickly shifting to being one of federal policy That you can you can get the the federal government to set standards for how for um, what uh, what in a surveillance system should do and what it should not do But if you're raining money down and you're not controlling how it's spent the fastest thing That the federal government could do is just turn off the spigot or say that that this can't be used for this kind of technology until we study it more Um, but I'm sure the vendors love the money, you know fallen out of the sky that basically the way it is now And scott earlier you mentioned to like did we think do we think there would be any sort of pushback against regulation? I know we didn't address that specifically But I I would personally think that that's likely because uh shot spotter as an entity has retaliated to various other things that have happened Um, they currently have a lawsuit with vice after vice reported on some of their misconduct Uh, and so we we've seen things like this Uh happen where they what we think right like I can't make this claim 100 But it looks like they try to produce a chilling effect on news news groups that give unfavorable pr Um among other things as well Okay, um if anybody else would like to ask a question feel free to turn on your mic Feel free to turn on your camera or just put it in the chat at this point in time We're kind of in the open question phase by now larry's got his hand raised so Yes, thanks. Yeah, go ahead Thanks. Um, I'm wondering if the Installation of shot spotter is the same everywhere in other words Uh Do they all retain records? Uh, the same way Uh, does the system pick up? Uh Voice sounds Within the range of their microphones that kind of thing do they all the all the systems look the same? Yeah, I can take that so, um, I think basically yes, so The the system is deploying the microphones the sensors how that interfaces with their algorithm. That's always a shot those are shot spotter sensors. Um, so yes, um The way that it interfaces with the the local police's dispatch and That that may depend on what dispatch system exists in the city so like in chicago the shot spot alerts go to these um local like real-time crime intelligence rooms strategic decision support centers what they're called that are in each police district and And they send the police officers out in some cities. I think it just goes directly out to police officers on like smartphones or, um computers in their cars so that that That interface I think varies depending on the police department in terms of like the listening and recording That again, I think is uniform Each sensor is recording 30 hours if you're interested in like the privacy side of things, um, yes, there was an audit done by the policing project And why were they examined this sort of data retention? practices of shot spotter and my impression is that shot spotter doesn't want to be in the audio surveillance of voices Market, so they want to run away from that that they've gotten significant criticism about the idea that they're basically wiring up neighborhoods for um with microphones for audio that could be used as like audio surveillance and they want to Focus on the use of their system to detect gunshots. Um, that said, um There have been a couple cases where voices were picked up on sensors and We're used in trials in those cases. It was because the voices were um close to the sensor and also Within just a few seconds of the supposed gunshot so it's like in the time period before and after so it's part of the audio But I'm not aware of any cases where police have tried to pull some of that 30 hours of data off of a sensor to hear the actual well spoken but They they certainly could you know shots butter can go and pull that audio off of their sensors if If police demand it This would be a question of what kind of the court order or subpoena or warrant or something but Yeah, hopefully that's helpful Uh follow up, uh, so that would mean that a retained conversation Could be used by the police to go back with a warrant Even though they had no original suspicion Well, so the the the way I think it's it's theoretically possible, but the the way that the System works is that police don't Have access to the audio like by default or through a portal The audio just stays on the sensor unless it's allowed pop or bang or boom or whatever which automatically gets sent out to shot spotter or if shot spotter reaches out essentially to the sensor and Pulls pulls the audio off if the police make a request after the fact And I'm not aware of a situation where the police have done that basically called up shots spotter and said Hey, we think something went down on this corner where do you have any audio so we can hear people talking? I don't know that happening If it's theoretically possible it also has occurred to me that if somebody you know happens to live In like an apartment building like right next to a sensor or something um, are they inadvertently being Uh, you know eavesdropped, um You know when they're on their balcony or something, um, that certainly could be happening, but again not aware of any examples Thank you Yeah, I think there's a fourth amendment issue here that really hasn't been tested in court But the problem is that the fourth amendment doesn't really apply to private parties It only applies to the government, but then you have the the surveillance vendors that are acting Kind of as an armor and extension of the government And so it's it's one thing if you're I mean It's one thing if you're driving a delivery truck and you happen to see crime and you report it And that you're acting as a commercial entity But you just it's sort of an incidental reporting And then there's it's another thing when you're kind of joined at the hip with the police And you've got a data terminal right in the police station and and maybe right in the fusion center And so we get into a cycle where it's search warrant search again Kind of situation where the first search is being done by a private party But they're so deeply connected with the police that it's kind of a bypass of the fourth amendment And I I don't think this has really been tested in court, but that's essentially where the problem is Um, I see that chuck has his hand raised. Do you want to did you want to jump in? Yeah, I I just did that because I wanted to jump in at some point I wanted to just mention In kind of reply to what larry had said In a more to the average person To somebody on the street that we would be speaking to I think the impression That the technology provides is that you you when you understand that it's on all the time But it's waiting to be signaled by a loud boom or a pop or something like that And that's the kind of the indicator of what will let They'll know that they need to analyze something It it reminds me of being at the city council Uh a few years ago with scott when we were listening to them talk about body cam footage and whether Whether the microphone and the camera was on or off and you should have seen the average city council person the average human listening to The explanation of that well, it's technically always on but we don't actually know What this is and explaining that they can only retrieve 20 seconds before Something happens that it you weren't supposed to have known that it had happened because it wasn't recording Except that it's always on and it is really that confusing and uh for me What is most concerning is the fact that whether You know, they can actually extract that and use it Somehow it I'm more concerned about the fact that the sensors are there and that they will be fed through a fusion center And we don't really know how that's retained and whether it will can be Used in the future in some way Like alexa exactly as karim is saying And so because we don't really have any understanding of what people are going to be doing with that Stored information. I mean it's electrons. I guess is the way people describe it It doesn't take up space so you can just store and store and store and store What will the algorithms of the future be looking at and listening to because of this stored nothingness that was supposed to be nothing That's how I as an average person look at it And that's why I I don't like it and I don't want it there because I don't think we need to have sensors everywhere And I have lived in an apartment and I don't want to be on my balcony and have that exact scenario that jonathan mentioned happened to me Yeah, and I mean to your point chuk it it sounds like I always get the sense that one data is collected in that way Um, and it's used to profit these entities these private parties in particular It's uh, it feels kind of shitty, you know, like at the end of the day It's you are sort of producing this uh valuable product, which is your data and without any sort of uh Uh real form of consent over using it. It's just used to profit these sorts of systems so it um, it sucks reflecting and saying that like This data or these systems can remain active and produce almost print out money off of stuff that In you know in a sense really belongs to you or is yours? We're right. We don't know who who retains What for how long but we do know that super computing exists and that Saving whatever data is there is no problem for the human race at this point and it's a matter of What would you do with it at a later time? And uh, they unintended consequences after the fact Are not the kind that we need to get to involve them without thinking about it ahead of time Thanks for letting me participate Thanks Chuck. Um Keith you had a comment Yeah, I'm looking at this from an engineering perspective If I've heard multiple constraints one is Uh The technology isn't trained on actual gunshots. It doesn't actually work has lots of us positives There's fourth amendment concerns with the way it records audio where it's deployed Is along racial lines partially that's also What they're going to come back with on that is they're going to say That's where statistically the crime is and we have a limited budget to where we can put it Um, there's also the concern of hey, this the system's expensive for what they're getting and there it doesn't It does deal with responding to crime But it doesn't do anything to work on the social issues that result in crime So it how is there someone we can do both? Well, looking this is an enduring problem if I Make a supposition that we could invent the perfect ai that can get five nines accuracy of gunshots did not record Speech audio at all and didn't save it anywhere So it's not possible for the government to use it in some nefarious way and it met all the Fourth amendment requirements. The system was deployed ubiquitously regardless of Population in the area or crime rates. It was just everywhere And its cost was free. So it doesn't impact on spending money elsewhere Would that in fact be something that we want would it actually produce the results? That that the shot spotter system is supposedly intended to do which is Notifies the police when there's a gunshot accurately so they can respond quickly and if there's an event that there's people they're injured They can respond to them Well, that's actually even if all of that's stipulated that's still a maybe we're not sure And the other problem with with their argument with shot spotter is when it comes to that perfect system The ai that's five nines doesn't meets all the fourth amendment requirements. It's inexpensive It's ubiquitous because it's inexpensive That's waving a technological magical wand Because that can't actually be achieved not yet and it's potentially a could I've not seen any research papers in the universities yet on Shot spotter technology that not just shot spotter technology, but the whole idea of machine learning to detect gunshots That's questionable As to whether that can even be done accurately. I know that from Doing wi-fi analysis which wi-fi by the way reflects off buildings reflects off trucks going down the street just like audio does Makes direction finding what's called fox and hound of wi-fi access points at conferences in downtown Atlanta virtually impossible Even with direction highly directional antennas because of reflections off of buildings So there's there's some question as to certain environments to whether it's even possible to do This given a magic technical wand of good ai that could recognize that could even triangulate properly so given all of that And then throwing into the mix that what shot spotter is vending is none of that It's a non starter that they've they've also their lips are moving their line So when I take all of that into account my take on it is no That technology is is not ready for prime time because it doesn't meet that that requirement And it's robbing money from where else it could be spent better And I don't I'm not willing to help them now if they want to contract me as an engineer to help them Review and make their product better fine, but I'm not willing to do their homework for them for free. Sorry Uh a lot of I have a lot of great points there. I thought that was really, um Useful I mean in terms of like the physics of it you're right like the the You know sound propagates in complicated ways that depends on reflections on buildings apparently depends on like the temperature of The air which affects the speed at which sound propagates wind can affect the Speed at which the sound propagates. It's um Uh, you know, it's like a complicated business. So Uh, I think that there's like questions about the fundamental reliability of the The methods for triangulation as well as you know distinguishing whether something is a gunshot or not, um The first place so Uh Yeah, all those points are really well taken And I guess another thing to think it was just at like a fundamental level Even in a perfect world. This is A system that is at best getting police out to the location where gunshot was fired Some minutes after it occurred and you know, it doesn't take a rocket scientist to figure out how you Um avoid getting caught by police where you fired a gunshot Like you leave quickly so so so, you know, um like Uh, even if it were magic and 100 accurate and stuff. I don't Know that it would actually be a very effective pool So Uh, and the you know the proofs in the look shot spotter has been marketing the system and tweaking the engineer for 20 years Um and yet The data on the ground consistently shows that police find no evidence of a gunshot Somewhere between 5 and 20 percent of the time. Sorry. It's number between 95 and 80 percent of the time When they arrive at the scene so you know Kareem had a couple of questions that are back in the chat. They've scrolled away, but Um, here's one have the feds used any shot spotter data from a local law enforcement entity in their investigation Yeah, they they they have it doesn't comes up sometimes. Um, it doesn't seem like it's that frequent but um they They have I put in the chat a case. It's called rickman That went up to the seventh circuit and in that case the person I believe the person was stopped after a shot spotter alert And the case kind of discusses briefly like The status of the shot spotter like how how to treat the shot spotter alert for fourth amendment purposes It's doesn't analyze in much depth. But basically says like Maybe we should treat this like an anonymous tip. We can't really tell if it's reliable or not That's the analogy that they made in fourth amendment context If you're interested in that like in the connection to stop and prison the fourth amendment we published Published the filed an amicus brief basically arguing that shots spot alerts aren't sufficiently reliable to justify police stops And also that the information that a shot spot alert generates Isn't specific isn't specific enough to a person to Establish grounds to stop a person so Um So in the absence of like other facts that the police officer encounters at the location that they are sent out to Police shouldn't be stopping people in response to shot spot alerts And I can drop a link to that in the chat if you're curious We made an analogy to police Police dogs that sniff for drugs So for police to use dogs that sniff for drugs. They have to actually like Do a training program with the dogs and show that the dogs are generally reliable To some standard in a controlled environment, there's no analogous testing of shot spotter at least not it's been published. So So we thought we think that's fatal to using it as basis for a fourth amendment search Yeah, I wanted to ask are there any more questions. I think we're getting close to time I'm going to need to stop the recording pretty soon and but we can still hang out for for casual chat Any any any last or final questions? Okay, uh seeing none. I'm going to go ahead and start Wrapping up. Uh, I wanted a huge thanks to Allie Jonathan and alex for this. Uh, what's really changed in this fight Is that now the activists in different cities are starting to talk to each other Um and starting to coordinate more and that just wasn't happening every city was a was a Kind of a tower unto itself And so now that we're connecting we're able to connect Uh, I think we've been able to connect for a long time But but we may actually have COVID to to thank for kind of getting our things together Getting us together in a sense because we haven't been able to meet people locally as well But since we're um since we're meeting online We're able to jump over geography more easily. So That's actually helped in a way for us to be able to get together from different cities So it's a it's sort of a blessing in disguise. I guess in a sense um But yeah, I I really want to thank everybody who's attended tonight We don't normally have weekly meetings, but we will actually have another meeting in a week to talk about state legislation and two bills in particular that look like they're going to be regulating social media At the state level They would regulate much more than social media But I think they were inspired by social media and they were inspired by regulating big tech companies So we'll talk about that Next week it'll actually be the same url Same uh day of the week wednesday and same time seven eastern So I want to thank everybody who's uh Who's uh come into this tonight and thanks for sharing your concerns And I'm sure this is just the beginning of the conversation and it's kind of ongoing I'm going to go ahead and uh shut down the stream and and the recording and once the you see that the recording stopped We can start the um we can start the more casual participation Uh, but uh once again, thanks everyone Thank you so much. Thank you