Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Canada AI Technology

Police Across Canada Are Using Predictive Policing Algorithms, Report Finds (vice.com) 158

Police across Canada are increasingly using controversial algorithms to predict where crimes could occur, who might go missing, and to help them determine where they should patrol, despite fundamental human rights concerns, a new report has found. From a report: To Surveil and Predict: A Human Rights Analysis of Algorithmic Policing in Canada is the result of a joint investigation by the University of Toronto's International Human Rights Program (IHRP) and Citizen Lab. It details how, in the words of the report's authors, "law enforcement agencies across Canada have started to use, procure, develop, or test a variety of algorithmic policing methods," with potentially dire consequences for civil liberties, privacy and other Charter rights, the authors warn.

The report breaks down how police are using or considering the use of algorithms for several purposes including predictive policing, which uses historical police data to predict where crime will occur in the future. Right now in Canada, police are using algorithms to analyze data about individuals to predict who might go missing, with the goal of one day using the technology in other areas of the criminal justice system. Some police services are using algorithms to automate the mass collection and analysis of public data, including social media posts, and to apply facial recognition to existing mugshot databases for investigative purposes. "Algorithmic policing technologies are present or under consideration throughout Canada in the forms of both predictive policing and algorithmic surveillance tools." the report reads.

This discussion has been archived. No new comments can be posted.

Police Across Canada Are Using Predictive Policing Algorithms, Report Finds

Comments Filter:
  • by will_die ( 586523 ) on Tuesday September 01, 2020 @12:20PM (#60462010) Homepage
    So the police are looking for higher crime areas and then making sure that police are in those areas in hopes that the increased presence will prevent the crime and if not they will have an increase chance of catching the criminals.
    So what is controversial about that?
    I'm in a low crime area and if police want to station a cop car or two all the time around here feel free to.
    • I think its probably eluding to things assocuated with higher patroled areas, such as racial profilng and unwarrented road checks, which has been in the news quite bit lately in my area.

      But more generally, the big issue compounds the problem of blanket servalence, blanket proccung it. Its great I guess if it prevented child abductions, but I worry the probable effect may result in mailed out j-walking fines because somebody caught you in a frame of a picture they posted on facebook.

      • It's not really that, but a question about what an algorithm would do assuming the data you fed it had some bias that isn't accounted for. It's not present over-policing, but past over-policing that results in issues where the algorithm just repeats the same bias that continues to produce the same results all over again in a self-fulfilling prophecy kind of way.

        There's some merit to the argument being made, but I don't put as much stock into it as those who usually present it as their reason for opposing
    • by lessSockMorePuppet ( 6778792 ) on Tuesday September 01, 2020 @12:40PM (#60462096) Homepage

      I'm in a low crime area and if police want to station a cop car or two all the time around here feel free to.

      To justify that, they'd have to arrest some people, then you'd eventually live in a "high crime area". Seriously, what's so difficult to grok about selective enforcement?

      Cops aren't sitting around patrolling rich neighborhoods looking for Mr. C Suite's coke dealer. Must only be a crime if you're poor enough to be a viable target for bullying police.

      • Like it or not, that coke delivery in Mr. C Suite's neighborhood isn't going to result in a drive by shooting or gang violence. The reason some people are able to commit crimes their entire lives and not be prosecuted for it is because they are otherwise able to behave themselves, not draw attention to themselves, and not hurt others.
      • by sdinfoserv ( 1793266 ) on Tuesday September 01, 2020 @12:51PM (#60462158)
        "Mr C Suite" hasn't had a coke dealer since the 80's. Poor neighborhoods have higher crime rates because there's more crime. You can argue over why, like, "De-Industrialized" zones have no jobs forcing people into the underground "aka illegal" economy, but the reality is poor areas have higher drug use, higher incidence of mental heath problems, fewer safety nets, less access to nutritional food, lower educational levels - all these contribute to higher crime. Where there is crime, predictive analysis puts cops. now it also happens that a higher percentage of people of color are poor.. that part really gets the "you're attacking people of color" crowd worked up, but that's not the picture. Look at Chicago Ill for example. It's roughly 1/3 white, 1/3 Latino, 1/3 African american by population numbers... Yet, 80%+ of gun violence is black on black... Why is that? Why is even asking the question considered racist? If you can't ask questions, you can't get solutions. Time to stop being sensitive and look at real cause / effect then decide if we as a nation are willing to do something.
        • That seems inaccurate. Lots of rich people buy drugs, and they source their contraband from - drumroll please! - poor neighborhoods. Distribution networks tend to rely on poverty-stricken areas since the residents IN GENERAL tolerate more of such activity and/or have less power to successfully report it to authorities.

          If you have a major distribution point in the suburbs with people in various vehicles coming to and fro 24/7, yeah the locals will complain and get it stopped. In the hood? Nah. Popo might

          • I mean have you been involved with buying drugs? Or rich people buying drugs?

            Like everything else the rich do, they will pay premiums to avoid hassles. I knew a rich woman 30 years ago who would pay about double the going rate for pot. I could supply her and take care of myself for free. And I didn't even have to deal with scary people.

            And it's not like I had some great entrepreneurial idea -- there's a whole universe of people who have figured this out, and rich people can get tons of drugs delivered t

        • Poor neighborhoods have higher crime rates because there's more crime.

          That may or may not be an accurate statement. I can make crime rates higher in poorer communities simply by ignoring crime in rich neighborhoods. Any kind of statistic can be manipulated. You need to show to me that these numbers are accurate before I will take them at face value. I am not saying that the numbers ARE wrong, let's be clear about that. There are a lot of reasons to suggest that such numbers could be correct. But when you focus all your time and attention at problems in one area you may

          • by tsqr ( 808554 )

            I can make crime rates higher in poorer communities simply by ignoring crime in rich neighborhoods.

            Absolutely correct. If the police would stop ignoring all the street whores, corner drug dealers, and drive-by shootings in Brentwood, the crime rates in that neighborhood would skyrocket and crime rates in South Los Angeles would plummet.

            Cops go after the low-hanging fruit first, just like most every other field of human endeavor. You can call that "being lazy", "efficient use of resources", "the police chief is running for re-election", or "getting the most bang for the buck" depending on how you want to

        • by Xenographic ( 557057 ) on Tuesday September 01, 2020 @03:47PM (#60462904) Journal

          > If you can't ask questions, you can't get solutions.

          If you haven't noticed, they're opposed to literally every way of making police more effective no matter what. I mean, why else did we see convicted pedophiles, wife beaters and thieves caught violently protesting the police in Kenosha (and illegally bringing weapons there)? This is like foxes voting against more guards for the henhouse at this point.

        • - all these are correlated with higher crime

          FTFY.

          It's a chicken or the egg conundrum. Is the fact that these areas are lacking basic amenities and social safety nets the cause of high crime or are these areas lacking basic amenities and social safety nets because no one wants to move in, develop, pay taxes and invest in those neighbourhoods because they have high crime?

          It's quite possible that it's a mixture of both.

        • Why is even asking the question considered racist?

          A large portion of all language is context. The context of the question you ask is what determines if you're racist. Never analyse English in isolation.

        • Supposing your numbers are correct, it's not those who discuss them who are racist, the criminal gangs are. For sure if they allowed criminals from all colours to take part equally, it wouldn't be black on black killings.
    • by Ed_1024 ( 744566 )

      I had similar thoughts. Would people rather police were dispersed randomly (or even equally per head of population) or concentrated where they thought they were needed most? It seems that the process of guy-in-charge looks at map and decides where to deploy his limited assets to maximum effect has been semi-automated.

      I am sure the officers themselves would much rather be rescuing cats from trees out in the sticks as opposed to intervening in gang wars or drug deals...

      • by slapyslapslap ( 995769 ) on Tuesday September 01, 2020 @12:50PM (#60462152)
        There is not a small number of people who have committed themselves to the idea that all socio-economic groups and neighborhoods have the same amount of crime, and that it's only racist law enforcement that drives statistics to show otherwise.
        • by Shotgun ( 30919 )

          Those people suck at math. Because, all you have to do is count the dead bodies to know that it ain't true. Regardless of how you feel about it, dead people don't lie.

      • Even better what would the news be like if the police refused to answer a call in an poor area because the mandatory equally distributed policing had allocated to a low crime area. Sorry mam your rape cannot be attended because your area is over policed, we are busy trying to find a rich white man to arrest to equal out the stats, try again next month. Of course if you want to really be equal it should be a white woman, because you know if the distribution of as statistic is not equal there must be some dis

    • Re: (Score:3, Informative)

      In the United States auto and property insurance companies have tried occasionally to use algorithms to set rates that are most fair to everyone. Objectively, "most fair" is not the same thing as "equal." So, invariably rates were higher for groups that factually represented greater risk to the pool of insurance consumers (largely based on real data from claims paid), and also predictably these groups tended to be represented by a single or small number of demographics. In all cases like this the consume
      • That's exactly how insurance rate pricing is supposed to work.

      • Usually the ones that sue and win are outliers within the demographic. You can have a black man with the same credit indicators as a white woman, and the man will come out with the lower credit score, get denied access to credit, and other "fun" stuff.

        It's hard to make computer algorithms that don't wind up being incredibly racist.

    • by Gravis Zero ( 934156 ) on Tuesday September 01, 2020 @01:01PM (#60462200)

      So the police are looking for higher crime areas and then making sure that police are in those areas in hopes that the increased presence will prevent the crime and if not they will have an increase chance of catching the criminals.

      So what is controversial about that?

      The problem isn't the algorithm, the problem is with the police. Police have been known to make stupid arrests for nonsensical reasons anywhere they are.

      When you combine this bad policing with the algorithm it results in a feedback loop. Stupid arrests lead to higher presence leading to more stupid arrests. You end up with a population that resents the police and therefore are tempted to actually commit crimes against them.

      This can be summed up as, "the beatings will continue until morale improves".

      • And when the police stand down the situation quickly descends into complete lawlessness as we see every night right now. Instead of demanding the police be impossibly perfect have you considered demanding that people actually abide by the law? Or is that too racist?

        • Instead of demanding the police be impossibly perfect

          If they can't not murder people then they shouldn't be a police officer. If they are easily intimidated and think a gun is the solution for all situations then they shouldn't be a police officer. If they feel like the whole world is against them and they have to use heavy handed tactics then they shouldn't be a police officer.

          These aren't impossible demands, it just seems that some people aren't cut out for the job.

          • Democratic governments and democratic sheriff's still don't seem able to find cops meeting your standards ... regardless of race.

            Police in black majority areas have to survive a couple decades worth of street duty without serious personal injury on average. They will do whatever it takes, whether it's not policing in the way you like or alternatively just not policing at all.

      • The problem isn't the algorithm or the police. The problem is that the results are politically incorrect. This leads to people claiming that they are being targeted over race, when in fact they were being targeted over hard data showing crime is more likely in certain areas.

        The computer doesn't know or comprehend what race is. It simply knows that in certain areas you have increased rates of certain types of crime. Depending on how sophisticated the models are it can perform correlation with dates, seasons,

        • The problem isn't the algorithm or the police.

          I literally just explained the problematic interaction of both of these things. Your response is does not address my claims.

          This leads to people claiming that they are being targeted over race,

          I never mentioned race in my reasoning, so why have you brought up the subject of race when it has nothing to do with my argument?

      • by Jack9 ( 11421 )

        > The problem isn't the algorithm

        > This can be summed up as, "the beatings will continue until morale improves".

        Creating a disconnected narrative when you have already pointed out that the topic is a non-issue. Classic troll.

        • Creating a disconnected narrative when you have already pointed out that the topic is a non-issue. Classic troll.

          Who said the algorithm is a non-issue? I said it wasn't the problem but that doesn't mean it doesn't cause problems for an already broken system. It's like a 25 Amps socket connected with wire rated for 5 Amps and here you are saying "pulling 25 Amps is a non-issue". If the system were correct, 25 Amps wouldn't be an issue but the way it is now, you're going to burn down your house.

    • by Shotgun ( 30919 )

      I'd say they can park the police car at the top of my driveway, but I've got two neighbors that are policemen, so someone else deserves the love.

    • by Jhon ( 241832 )

      "So what is controversial about that? I'm in a low crime area and if police want to station a cop car or two all the time around here feel free to."

      I can't speak to Canada or other areas -- but I can speak about Los Angeles.

      Basically, yeah -- you have the first part -- they look at crime stats (CompStat in LA).

      The idea is good. Use computer analysis of 10 years of crime data to try and predict where and when crimes were more likely to occur based on past and current crime trends. The computer would then

  • Nobody ever asks (Score:5, Insightful)

    by slapyslapslap ( 995769 ) on Tuesday September 01, 2020 @12:59PM (#60462194)
    the people who benefit from this type of police presence and enforcement what they think. People who live in poor crime ridden neighborhoods really don't like it when they become crime victims, sometimes repeatedly. People want to "defund the police", but they aren't the people who require their protection.
    • by eepok ( 545733 )

      Exactly. I grew up in a poor, high-crime area. I wish there were MORE cops there. Yes, it was a majority Mexican-American neighborhood, but it was also a majority GANG neighborhood. Have you ever had to leave someone home while everyone else went to visit grandma because if it didn't look like someone was home, the place would be burgled? I did. That was life. It sucked.

    • The idea of "defunding the police" is to reallocate those funds to ways to proactively prevent the conditions which create crime. Instead of sending a cop to stop a domestic dispute, send a conflict resolution counselor. Instead of cracking down on drug dealers, invest in addiction treatment. Instead of sweeping the homeless guy with mental health issues out of the suburbs and putting him under a bridge, get him in a mental health program.

      In other words, treat the root causes, instead of ignoring them un

      • That is the version of "defunding the police" that sane liberals have tried to salvage from the original version called for by the far left, which literally wanted to abolish police. Liberals need to stop pretending they don't have a problem with radicals too.
        • Isn't that a good thing though? People took an idea that was too radical, made something pretty sensible out of it, and the majority are now backing that idea.

  • We've already seen that demonstrated over and over again.
    'Algorithms', having no capacity to 'think', critically or otherwise, not 'knowing' anything whatsoever let alone 'right from wrong', should not be used to determine anything whatsoever with regards to human interactions. They are not better than us in any way shape or form, they are a mirror for the worst we can be, and as such are useless.
    The more I see and read about so-called inaptly-termed shitty 'AI' the more useless I think it is, the more o
    • Police need to be boots-on-the-ground in their neighborhoods, they need to know the people who live in those neighborhoods, have a relationship with the people they are supposed to be 'serving and protecting',

      Oh, an olive branch. You actually want more positive interactions with police officers instead of assuming they're all garbage

      not sitting on their fat doughnut-fed asses staring at computer screens and being told what to do by shitty 'algorithms'.

      Oh, there it is. I wonder why they don't listen to you?

      • Oh sure because they're all perfect Angels who never do any wrong, never have any malicious intent, and all have IQs of at least 140 and the wisdom of an ancient philosopher-king.
        Stop defending LEOs who are shitbags. The system is profoundly broken. I'll call out fhe bad ones all I like; does that make you angry? Good. Enough with the complacency.
        • Oh sure because they're all perfect Angels who never do any wrong, never have any malicious intent, and all have IQs of at least 140 and the wisdom of an ancient philosopher-king. Stop defending LEOs who are shitbags. The system is profoundly broken. I'll call out fhe bad ones all I like; does that make you angry? Good. Enough with the complacency.

          I'm just trying to reconcile the way you describe law enforcement with your statement

          Police need to be boots-on-the-ground in their neighborhoods, they need to know the people who live in those neighborhoods, have a relationship with the people they are supposed to be 'serving and protecting'

          If you called me fat, lazy, malicious, stupid, and much worse, I would have no desire to get to know you or to come anywhere near you unless I had to.

          • Re: (Score:2, Offtopic)

            Oh gee I'm sorry I don't spend half my waking hours writing a few thousand words and having it all vetted by a dozen proofreaders and fact-checkers so I only insult the specific cops that need insulting! My bad that you don't realize that painting with a broad brush is necessary. But I'm not apologizing for that in reality and I'm not walking any of it back either. If you're unhappy with what I had to say then that's your problem. There are plenty of shitty cops in this country, in this world we live in, an
    • Algorithms process data. Therefore algorithms are racist and bigoted whenever the data or the actual processing is racist and bigoted. I think we can safely exclude the latter, programmers have enough problems with, you know, programming, to further complicate it with 'racist code". So the data, which is usually about real world, is racist and bigoted. So real world is racist and bigoted. A bit harsh, but what do I know...
    • by Shotgun ( 30919 )

      They do that instead and they'll know better than any shitty 'algorithm' could ever tell them what they need to be looking out for.

      If they do that, will they be allowed to defend themselves when a guy they are trying to arrest for sexual assault goes for a knife?

  • It is a very different story to use such things to find and protect potential victims than it is to find and investigate potential suspects.

    Huge difference between looking into a child to see if they are being abused and looking at the 60 year old single, unemployed guy to see if he is abusing a child.

  • Used to determine the crime rate in an area, I don't see a particular problem. I do see great potential for bias if it is used with regard to individuals, either as potential criminals or victims.

    the above says "Right now in Canada, police are using algorithms to analyze data about individuals to predict who might go missing"

    That sounds like it is being used to make predictions about individuals. I see that as very concerning because past bias can be baked into the decision algorithms. leading to more sel

  • I live in Canada.

    And to the best of my knowledge, they aren't actually arresting anyone who hasn't actually done something wrong here. While it's probably true that increased police presence might lead to unwarranted questioning of innocent people by police, in my own experience with such matters (I have, for what it worth, been questioned by police without any reason or provocation whatsoever on my part I don't know how many times in my life, especially while I was a young adult), just politely and truthfully telling the police what you are doing at the place where you happen to be can turn what might initially otherwise feel like a completely unwarranted interrogation into a quick two minute conversation with the police before they thank you for your cooperation and you can carry on doing whatever you were doing before.

    Yeah, I understand that it's perhaps less than ideal when you feel like you are being singled out because of factors that are wholly beyond your control when you have done absolutely nothing wrong.

    But the fact of the matter is that an increased amount of police presence in a community *does* reduce the amount of crime in an area.

    If they actually start arresting people who haven't committed any crime, then obviously that's a problem, but just as it's inappropriate for them to arrest people who have done nothing wrong, it's equally inappropriate for anyone else to presume that such inappropriate behavior by the police would somehow be inevitable before the actual statistics are in.

  • Let's stop and search that black dude, he looks like he's up to no good.

  • Get rid of them. Get cops who are NOT RACIST, BIGOTED, or otherwise BIASED. Also not prone to resort to violence as a first go-to.
    Tough shit if you don't like what I'm saying. Things need to change. Starting NOW. Scream all you want, IDGAF.
    • Tough shit if you don't like what I'm saying.

      What you're saying sounds emotional, not rational.

      Hardly anyone likes racist cops. We all oppose it. Who are you disagreeing with?

      • Hardly anyone likes racist cops. We all oppose it. Who are you disagreeing with?

        The racist assholes who would SILENCE me and people like me, because they'd rather see non-whites continue to be targeted.
        Do you really think this issue is going to go away in a news cycle or two? Are you one of those people who read the news and go "Oh, wow, some other black kid got shot by cops, how horrible!" then turn to the Sports Section and forget about it? This is NOT GOING AWAY ANYTIME SOON, not until there is real change, and not just lip-service. Enough is enough.
        I'm not some kid, I've been aro

      • What you're saying sounds emotional, not rational.

        Don't you get it? The time for 'polite discourse' is long since past. IT ACCOMPLISHED NOTHING. They TRIED 'polite discourse'. They TRIED 'working within the system'; none of it worked. Why? The system is corrupt and biased against them. They can protest and make all the noise they possibly can, disrupt everything, BE HEARD, or they can continue to spin their wheels, be led around in circles by 'The System'. Which do you think is going to force change at thi point?
        Meanwhile OF COURSE the racist types WANT t

        • Who are you talking to? We have an agreement that racist cops are bad.
          • Seem to have missed a close-boldface tag in there. Sorry about that. xD
            You wanted to know what I was on about and seemed to think I'm not 'being rational' about this. I wanted to demonstrate to you that while there's an unavoidable emotional component to this for everyone. there are definite issues that cannot be ignored anymore. But, yes, I'm passionate about this. Tired of this country, which has the potential to be so awesome, sliding farther and farther into the muck.

A computer scientist is someone who fixes things that aren't broken.

Working...