Worcester doesn’t need “crime forecasting”
But the cops get what they want here
Today’s MLK Day, meaning it’s as good a time as ever to listen to this speech he gave in Worcester if you haven’t yet. At Temple Emanuel way back in 1961.
MLK, after recounting some of the progresses made on the racial justice front up until that time, said the following:
It would be a marvelous thing if I could stop my speech at this point. Bring it in and just take a seat. First it would mean giving a short speech and it’s always nice to make a short speech. It would also mean that the problem is solved in America and that we don’t have anything to worry about. But you see, if I stop at this point I would be merely stating a fact, and not telling the truth. You see, a fact is merely the absence of contradiction but truth is the presence of coherence. Truth is the relatedness of facts. Now it is a fact that we’ve come a long, long way. But it isn’t true because we’ve got to add the other side to it and if I stopped at this point I would leave you the victims of a dangerous optimism. I would leave you the victims of an illusion wrapped in superficiality. So in order to tell the truth, it is necessary to move on, and say not only have we come a long, long way, but we have a long, long way to go.
Here in Worcester some 60 years later, King’s words ring true, as we’re still going full speed ahead in our mission to avoid substantive changes to a racist system of policing at all costs. We’re just not going to go there, folks, don’t even try it.
It is a fact, if I may be so bold as to riff on King’s speech, that the institution of policing in America is rooted fundamentally in racism. But it is true that policing in America upholds the racism from which it was rooted.
I haven’t made a pitch like this in a while but trust me when I say no other journalist in Worcester is going to talk about the cops like this because they either won’t or they can’t so consider subscribing so I can continue to do this :-) thank you.
The City Manager over the summer promised a package of police reforms coming in the fall. It is now very much the winter, and that package must have got lost in the mail or something because where is it? Not like there was going to be much of anything good in there anyway. Just more ways to spend more money on the police without substantively reforming anything. Just look at the way the City Council and the manager dismissed with an annoyed wave of the hand the demands of the community to simply not increase the police department’s budget. They’re clearly uninterested. The Police Chief will not admit he’s ever seen any racism in his department. He’s not interested.
Instead, the City Council tomorrow will evaluate a proposal from the Police Department to spend about $150,000 (and then about $200,000 year over year) on an artificial intelligence software that “allows the command staff to direct patrols based on data-driven crime forecasting.”
I know we’ve been all Orwell this and Orwell that lately but really can you think of a more dystopian term than “crime forecasting”?
The program is called ShotSpotter Connect, an expansion of the ShotSpotter program the police department already employs. What ShotSpotter does is it picks up gun shots, or things that sound like gunshots, and it tells the cops hey, there was a gunshot over here, go check it out. What ShotSpotter Connect adds is an artificial intelligence component. From the chief’s memo in the council packet:
“ShotSpotter Connect uses computer aided dispatch and artificial intelligence-driven analysis to help strategically plan directed patrols and tactics for maximum crime deterrence—effectively extending the capabilities of ShotSpotter Flex to deter gun crime before it occurs. The system incorporates data such as crime events, temporal cycles, and geographic variables to produce high probability crime forecasts. These forecasts are then turned into patrol plans (“directed patrols” or “patrols”) that are available to commanders, crime analysts, and officers in the field. These patrol plans change for each shift and by day of the week.”
So what this computer program does is it takes all the data collected by ShotSpotter, as well as the time of year (the less douchey way to say “temporal cycle) and “geographic variables,” which could mean pretty much anything, who knows, and creates maps telling the police where to send patrols and when.
Would you want to live in the neighborhoods that an artificial intelligence has determined is more likely than others to experience gun violence, and therefore your neighborhood is more heavily patrolled by police, which necessarily creates more crime? I wouldn’t. And I probably do live in one of those neighborhoods. The police are doing just fine without some detached AI telling them exactly where to go. The implications of the use of AI in this way are heavy. Computer-determined pre-crime based on where you live. The cops come into your neighborhood expecting gun violence because the computer told them so. But while they’re there they might notice you’re driving an unregistered car. They might find someone with a warrant out for missing a date with his parole officer. They might see some kids doing wheelies in the street and decide to take their bikes. They might see a homeless person with nowhere else to go and tell him to go somewhere else, anywhere but there.
If they’re there, they’re more likely to do the routine bureaucratic things that cops do to poor people to keep them poor. And now, if we implement this new technology, the patrols will be “more efficient,” meaning there will be a higher concentration of police in more specific areas.
The cops have already done a great job over the past couple decades, or maybe even since they first existed, ensuring the “bad neighborhoods” stay bad and that they don’t expand over into the “good neighborhoods.” They’re plenty capable of doing the dirty work of ensuring there’s an underclass by way of routine economic and at times physical violence. They do not need the help of an AI. They do not need more “efficiency.” More from the chief’s pitch:
Directed patrols are automatically generated at the start of each shift. Top crime risks for particular areas are plotted as colored boxes on a beat map for the upcoming shift. Each patrol area measures 250m by 250m by default (size is configurable) and is color-coded according to crime type. Our Crime Analysts can add custom patrols or suppress recommended patrols as needed based on incoming intelligence and agency objectives. Connect provides useful insights to crime analysts, command staff, and patrol officers who need to track and measure how effective individual officers and teams have been in meeting planned objectives while proactively visiting directed patrol areas.
There’s a lot to unpack in the above passage. That the tool would also be used to “measure how effective individual officers and teams have been” is doubly worrying. What are these planned objectives? I say this knowing how ridiculous it sounds, but do we have any control over what those objectives are in this, our supposed democracy? It sounds a lot like having a computer say oh, Officer X hasn’t been writing very many tickets lately, you should encourage him to write more tickets. And then he goes out there and he writes some more tickets and puts more desperate people in even more desperate situations which may inevitably lead them to commit crimes. How’s that for forecasting!
And this is all taking for granted that ShotSpotter actually works, which is sort of an open question. The company is notoriously shady about sharing the shot data it collects, and some studies have found that officers responding to ShotSpotter incidents have about a 30 percent chance of finding any sort of gun.
Journalist Matt Drange published a long and substantial examination of ShotSpotter at Forbes back in 2016. This line in particular still rings with a bit of that truth MLK was talking about.
Despite its presence in nearly every major metropolitan area of the country, there remains little external validation of ShotSpotter from researchers or government agencies, which often guide decisions made by local law enforcement.
Tough words here as well from a study on shot spotting technology in St. Louis, and this was published by Police Chief Magazine, so that should tell you something. The call’s coming from inside the house, etc.
Results show that AGDS simply seem to replace traditional calls for service and do so less efficiently and at a greater monetary cost to departments. Given the tepid results in guiding police to the scenes of crime and given the hidden costs of these systems illustrated here, AGDS might not be well-suited for the audience the technology is marketed toward. High-volume agencies will likely experience substantial increases in their call volumes with remarkably little to show for it, at a cost that might have taxpayers questioning the logic behind the expense. While this technology can be useful, especially from an analytical point of view, it is difficult to see how agencies benefit from expensive technology that increases financial strain on departments with its only discernible impact being fewer founded crime incidents.
In the Forbes article, there’s a vague reference to ShotSpotter analysts having trouble with things like fireworks. Another loud, gun-like noise that comes to mind are cars backfiring. Not for nothing, ] the neighborhoods where cars are more likely to backfire and people are more likely to shoot off fireworks are kinda the same neighborhoods here in Worcester. But the AI doesn’t know that. The AI doesn’t care. It all factors into the decisions it makes about telling the cops where to go.
So yes, I think you get it by now that I do not think adopting this expansion of the ShotSpotter program is a very good idea. But I would give it a 90–95 percent chance of flying past the City Council tomorrow with very little in the way of opposition. Maybe Councilor Khrystian King says something. Maybe he even votes against it. But that’s one voice out of 11. The rest wouldn’t risk running afoul of the police department and more important the police union, on which they depend in no small way for electoral success. The cops get what they want in this city.
Also, for what it’s worth, the money for this is coming out of the city manager’s contingency fund for some reason. Buried deep deep in the agenda is this motion, which is usually passed in a lump vote with a whole bunch of other financial motions.
A contingency fund is supposed to be used for future, unforeseen expenses. It’s there in case of emergency, basically, and I would hardly consider this an emergency. Not only are the cops getting a new toy to put boots on necks more efficiently, they’re not even paying for it. The city manager is dipping into his emergency reserves to make sure the police department has a new AI to tell them where the bad things are going to happen.
Not only is the city manager doing this, he’s perhaps more importantly not doing anything to reform the department.
This program is going to cost us about $192,000 annually after the first year. That’s like four substance abuse councilors or homeless outreach coordinators. Think about how much more good four people helping the community would do than one AI telling the cops where the guns go boom.
Defund WPD is active on this issue and I’m told they’ll be making a showing at the City Council meeting tomorrow. I hit them up for a statement and they told me “Again, Worcester's government continues to make funding decisions that steer money towards the police budget at the expense of community services that reduce local crime via education, housing, employment, and health."
They’ll be calling in and they want you to call in as well, a position I back hard. For more information on that, head over to their site and sign up for email updates and follow their socials and all that. They’ll have information on how to call in, making it nice and easy for you and they even have scripts for you to read if you’re not a big public speaker.
If you liked what you read just now consider throwing me some cash or telling a friend or even giving your friend some cash to throw at me.
Also today in “I’ve never seen an instance of racism in my department,” the Worcester Police Department lost another round in a 26-year-old lawsuit two officers filed against the department citing racial discrimination in promotional practices. The judgement, if the city does not appeal, could be as much as $6.5 million. These aren’t even the officers that I wrote about a while back. This is an entirely different case! Sargent’s claim he’s never seen any racism in his department continues to grow laughably thin. Will it stop him? Probably not lol.
And the legal battle between the city and the kids who were force-marched down Main Street after a June 1 Black Lives Matter demonstration continues. On Friday, lawyers for both sides met for a conference on the case, and, most notably, the lawyer for the kids involved said the Worcester Police Department has still not handed over an internal investigations report on the conduct of the officers, which as we know involves shoving young Black girls to the ground, stomping on phones, calling protestors bitches and whores, and the use of pepper ball rounds, foam-tipped bullets and smoke and/or gas grenades. Some of the charges are getting dropped, some of the charges are staying, and I’m told the Zoom conference the hearing took place on was bombed for about half an hour by anonymous right-wingers posting Nazi stuff, pro-Trump stuff, racist stuff and porn. I’m not saying the anonymous right-wingers were cops but I’m not not saying it.
Also not for nothing, there’s a tuition strike going on at Clark right now and part of their demands are that they want the campus cops disarmed, which would be a really big step. The administration has so far said that while they might wiggle a little bit on tuition they’re most certainly not going to disarm the cops. I wish the Clarkies best of luck fighting the good fight and I may revisit the subject in greater detail later. Clark is an institution of higher learning and as such it ranks relatively high on my shitlist.