Slashdot Mirror
Chicago's Experiment In Predictive Policing Isn't Working (theverge.com)
The U.S. will phase out private prisons, a move made possible by fewer and shorter sentences for drug offenses, reports the BBC. But when it comes to reducing arrests for violent crimes, police officers in Chicago found themselves resorting ineffectively to a $2 million algorithm which ultimately had them visiting people before any crime had been committed. schwit1 quotes Ars Technica: Struggling to reduce its high murder rate, the city of Chicago has become an
incubator for experimental policing techniques. Community policing, stop and frisk, "interruption" tactics --- the city has tried many strategies. Perhaps most controversial and promising has been the city's futuristic "heat list" -- an algorithm-generated list identifying people most likely to be involved in a shooting.
The hope was that the list would allow police to provide social services to people in danger, while also preventing likely shooters from picking up a gun. But a new report from the RAND Corporation shows nothing of the sort has happened. Instead, it indicates that the list is, at best, not even as effective as a most wanted list. At worst, it unnecessarily targets people for police attention, creating a new form of profiling.
The police argue they've updated the algorithm and improved their techniques for using it. But the article notes that the researchers began following the "heat list" when it launched in 2013, and "found that the program has saved no lives at all."
The hope was that the list would allow police to provide social services to people in danger, while also preventing likely shooters from picking up a gun. But a new report from the RAND Corporation shows nothing of the sort has happened. Instead, it indicates that the list is, at best, not even as effective as a most wanted list. At worst, it unnecessarily targets people for police attention, creating a new form of profiling.
The police argue they've updated the algorithm and improved their techniques for using it. But the article notes that the researchers began following the "heat list" when it launched in 2013, and "found that the program has saved no lives at all."
Maybe the enterprise of saving lives shouldn't be put entirely on the police? Obviously they have a role to play, but when we are talking about prevention, other institutions also have a huge role to play.
For example, many killings are the result of mental health problems that are going untreated. Part of the problem there is that the necessary care can be expensive. So....let's do something about that. What does the government-funded health care landscape look like these days? And what about educational grants (NOT LOANS) for mental health practitioners?
There is also still a strong social stigma against seeking mental health. Nobody is embarrassed to say something like "My arm was broke so I went to see the doctor," but the moment someone utters the phrase "mental health" everyone thinks of him as crazy, weak, and pathetic. This is ridiculous, and we need to put more social engineering to the task of fixing that (for example, a lot more television and movies can include scenes and dialogue implicating that the popular characters are seeing mental health professionals...and the attitude is that this is just a given that normal people do this sort of thing on a routine basis).
There are, of course, also economic motivators for murder. If poor people are being driven to these extremes by poverty, then why isn't one of the richest countries in the world doing something to address that? Why do we continue to abide the existence of charities that spend nearly all the donated money on their own staff and get no effective results? Why aren't we making more use of proven-effective programs like microlending?
There is quite a lot that can be done, and the police can't be left alone to do it all.
The failed experiment, called Chicago, just doesn't work. We have all known that for decades
Having more real cops on the streets would be best. I know that the ghetto is the last place Chicago cops want to be. But, that's where most of their murders take place. To curtail murders requires more than a computer program that does nothing actually predictive, only data mining.
Enforce the laws on minor crimes, and major crimes go down. You don't have to be a hardass, or pick on anyone in particular, just enforce the common, everyday laws that help keep things working.
We know this works.
When people notice that nobody is enforcing the little stuff, they start assuming that they can get away with the larger crimes - and they're usually right.
The problem is that, after a few years of it working, everyone relaxes and thinks "hey, crime is down, we can slack off a bit," and it's okay, for a while. Then things start slowly getting worse again, and the "corrective measures" tend to be away from the policies that were in force a few years before, because "they stopped working."
The problem with predicting where to go to stop crimes is that many of the crimes in Chicago are gang related, instead of property related. Houses to be robbed don't move, but rival gang members can be found anywhere. Predictive algorithms assume fixed targets.
If there was a real crackdown on Gangs, crime would decrease for a while, but I think that too many bribes are preventing that from happening. It would be far better to legalize drugs, defunding the gangs.
Of course, as a privileged white male from the suburbs, I could be wrong.
So, I rob a gun shop in the next State, in the sure knowledge that the Chicago PD will act as my fence? Yah, that's a good idea.
Note, by the by, that many places in the USA (basically all of them, since Chicago has one of the worst crime problems in the USA, and some of the most restrictive gun laws) get by just fine without worrying so much about guns in private hands.
"I do not agree with what you say, but I will defend to the death your right to say it"
"Chicago's Experiment In Predictive Policing Isn't Working"
Oh, so predicting the future doesn't work?
Damn, who could have guessed that prophesying is a tricky business? Who could have foreseen that? :)
Just cruising through this digital world at 33 1/3 rpm...
Stop lying, although I know you find that hard.
http://www.sarsonline.org/resources-stats/reports-laws-statics
52% of rapists are white, 83.5% of the population is white. .62 for whites, 2.9 for non whites)
Therefore non whites are over 4.5 times as likely to be a rapist.
(the normalised numbers are
That is on top of the estimates that rape is estimated over 5 TIMES less likely to be reported in non white communities.
None of which is good, but those are the facts.
All rape is bad, very very bad. Misusing statistics to focus on the wrong people is just as bad - you are disrespecting
the victims by trying to use their suffering for your own political purposes.
https://www.theguardian.com/us...
None of which is good, but those are the facts.
I generally agree with your post and share a dislike for misinterpreted statistic. However I think It's also important to understand that the raw data itself however empirical is quite fallible and cannot be trusted as hard evidence, the gathering of raw data is often as mistreated as the analysis. So referring to them as "facts" (perhaps not what you really meant) borders on 2nd order ignorance in my mind, they are indicators that even after correct analysis are open to interpretation and should be weighted based on the source of data.
The summary leaves out some important information that would tend to blunt the hyperbole it's trying to drive home. From the article:
It stressed that RAND "evaluated a very early version" of the list, "which has since evolved greatly and has been fully integrated with the Department’s management accountability process." It also points out that "the prediction model discussed in the report is the very early, initial model (Version 1), developed in August, 2012. We are now using Version 5, which is significantly improved."
A failing grade on the performance of a four-year-old version of the software, (and a four-year-old set of policies and procedures for using same), is hardly a reason to get all hot 'n' bothered, when what really matters is how the program is working today. It's news, and it may be significant, but it tells us nothing about the current effectiveness of the program in question. There are valid moral, ethical, and possibly legal issues around whether such a program should even exist, and whether the police are the right ones to be managing it - but that conversation shouldn't take place in the context of a FUD-driven summary of an article based largely on very stale data.
'The Economy' is a giant Ponzi scheme whose most pitiable suckers are the youngest among us and the yet-unborn.
52% of rapists are white
No. 52% of convicted rapists may be white, but conviction rates vary widely by race even for the same crimes. A white teenage date-rapist is much more likely to "get away with it" than a black guy.
83.5% of the population is white.
Wrong. Whites are about 72% of the population. This isn't 1960.
None of which is good, but those are the facts.
Yup. Other than being wildly inaccurate and misleading, those are indeed the "facts".