Police embracing tech that predicts crimes

Originally posted by acmpnsfal
Reply to post by getreadyalready

 

There is no way that would work, there are too many factors involved when an individual is making a decision to commit or not commit a crime. You can look at an area that has a lot of crime and conclude with statistics that more crime will probably happen in that area. But you cant take millions of people and account for all the varibles affecting their lives currently and in the past to try and guess who will commit a crime. It wont work, people are too complicated.

 

Posted Via ATS Mobile: m.abovetopsecret.com

 

i don't think it is a good idea, but I do think they could have some limited success. Think about the things that typically lead to crimes, and think about the ways we have of tracking those things. I bet we could create a database of every child abuse victim, and everybody with prior convictions, and everybody currently filing for bankruptcy or divorce, or is in the database as having complaints for child abuse. We could cross reference all those things, as well as current address and current employer and current drivers license status, and we could get a system that would map out the most likely people and places for crime.

Imagine if the program got a hit on a previous child abuse victim, with a violent juvenile past, that had been accused of child abuse their self, with a drug conviction, and the program found out through employer and tax databases that they had just lost their job, and their spouse had filed for divorce, and their address had recently changed to a neighorhood known for drugs and crime. It might be a worthwhile endeavor to put a cop near there and see what develops?

Again, I don't like the concept, but I'm not going to pretend it couldn't work. It could absolutely work.

Here something like 95% of robberies and violent crimes are solved within 3 hours and 60% know the criminal. Not the best and brightest.Over 97% within 24 hours. Guy just got nearly 50 years for armed robbery. Robbed the store with a gun, beat the man behind the counter for no reason then ripped the cloths off a female customer while his helper grabbed a couple of cases of beer. They made it 6 blocks and gave up without a fight and without opening a beer. 17 and 20 years old the 17 yrar old gets out in less than 5 the 20 year old will be 50 before he gets a chance at parole.

reply to post by mikellmikell


I know a guy here in my town that used to work for me from time to time. He grew up in a trailer less than 3 blocks from a convenience store his whole life, and he started robbing that convenience store at about age 15. He is now about 30, and he has robbed it multiple times, been caught every time, served time, got out, robbed it again, etc.

Yes, not the brightest bunch. He lived his whole life there, some panty hose over the head isn't going to hide his identity, LOL!

Originally posted by Talltexxxan
We are already in the world of precrime "minority report" style policing.

What is the purpose of a speeding ticket? You haven't hurt yourself or anyone else. You haven't taken away anything from anyone else. Speeding tickets are a form of policing through PRE crime. I get fined (or even arrested) because I "might" hurt someone else.

There are a mayrid of precrime offenses that can get you as little as a slap on the wrist all the way to landing you in prison for the rest of your life. All for somthing you MIGHT do. Jaywalking, possesion of drugs, almost all traffic violations, and a theres tons more.

Everyone should (re)watch Minority Report and think about how our world today already resembles the movie. Minus the jet packs and flying cars ofcourse

edit on 9-7-2012 by Talltexxxan because: (no reason given)

they say if you have more than a certain amount of a drug, you have "intent to distribute". they have NO way of knowing what you intend to do. but they can definately arrest you for "having intent".

Originally posted by bigfatfurrytexan
reply to post by Jerisa

 

Utter BS. There is no freedom in not being allowed to fail before intervention.

I agree, this completely negates the concept of free will. Some people do change their mind and back out of crimes that they were contemplating.

reply to post by Jerisa


utter twaddle

the movie you allude to was " minority reports "

which - used a panel of phychic precognitives to predict that :

person " w" would commit crime " x" at location " y " at time " z"

the software you cite predicts that unknown persons are more likley to comit cetain crimes at certain times in certain locations

ie [ i appologies - as i use UK crime treds / stats - because i have first hand experience - and access to raw data / anecdotal experience ]

fights betwwen males 16~ 25 are most common between 2200 ~ 01:00 friday / saturday - in areas with large numbers of bars

dui offences have a peak saturday morning 07:00 ~ 09:00 - this is because peopele attempt to drive to work after binge drinking

pickpockets target rail stations during rush hours traffic

illegal " street racing " is most likley to coocur bewween 22:00 and 02:00

prositution offences are rare between 09:00 and 17:00

i could go on - but the treds are obvious -

statistics show that certain crimes are most likley at certain times in cerytain areas

the software you cite - simply collates that data - and indicates that police resources should concentrate in certain areas at certain times

its not rocket science

Well, for starters, my "minority Report" spoof obviously didn't tickle any funny bones. Meh, never thought I would be a stand up comedian.

As for the the article itself, like all things government, there is the good, the bad and the ugly. What I have found in my own personal experience is that the government/industry takes a simple idea that could very well be a good thing, then they poke, twist, butcher,etc it until it is only used for their own gain and not necessarily what is good for the people.

The reason I used a spoof of Minority Report was because, in theory (minus the psychs), it could have very well started this way in its inception.

Seems like common sense. I mean it's only smart to focus on high crime areas. Now that technology makes it easier and more precise is a good thing.

Originally posted by Jerisa
reply to post by Myendica

 

Not missing the point, just musing where this could lead to...

It could be used as an instrument of surveillance. Criminals could be apprehended while committing crimes and lives and property could be saved.

Originally posted by getreadyalready

Sounds like a solid program based on historical data and statistics. I've got something similar for the lottery, but it doesn't seem to work very good.

I don't think the lottery is legit, for about four real good reasons. Just can't prove it, and likely never will. However, we did notice a rather *ahem* odd thing about the numbers. One of the guys here was putzing around with a program to look for tendencies in the numbers as something to do whilst learning Objective-C (you really have to have some reason to write a first learning program, that was his), and he found something really odd. Essentially, the first thing he tried allowed his program to guess 95% of the draws ever done in the powerball lottery (the five numbers, not the powerball itself) within about half a million attempts. That doesn't sound all that great, but it ought to take about five million random guesses or so to get even one of them.

I thought it was a program fluke I couldn't find, so I wrote one in C++ using the same basic approach but without looking at his source code and could get over 75% of them in 200,000 guesses. Not per draw - for the entire history. I was getting some of them in 100 guesses.

Basically the thing uses a strategy that's rather simplistic to produce a guess at five numbers, and looks through the lottery history to see if anything's ever matched that guess. When it hits one, it flags that, takes it out of the history list, and logs how many guesses it took to match that one. You can get every one of them in about 3 million tries - there are some far outliers, and by diddling the parameters on the thing doing the guess you can get more immediate hits at the cost of pushing the last 10% way out. I'm still not sure why it works at all.

I don't see anything wrong with this, they are just trying to put their cops in the places most likely to need cops. That is exactly what they are supposed to do. It isn't like their interviewing kids and arresting people before they do anything wrong.

I think it might make the cops a bit over-reactive and suspicious about people who might have nothing more than a bad day at work or worries about a family member. Then when they find they've screwed the pooch, they'll sprinkle crack on the guy to cover their asses. I understand the appeal, I just don't trust the implementors.

Of course, you understand it's also a really good way to get to use some of that NSA data you can't easily use otherwise, no? If they have some reason to believe you are a miscreant, the NSA will rarely inform FBI much less any lesser LEO that something's going down, and they won't/can't appear in court to introduce the evidence and provide expert testimony on its provenance and accuracy under oath. Well, not in most circumstances - there are a few exceptions. Generally when that sort of thing comes up they toss a hint over the wall the FBI to have LEOs or SAs at certain places and times around certain individuals, and lo! it just so happened that special agent Mike happened to be there and look, the guy did this really suspicious thing in front of me, giving me a plausible reason to open a case on him, and guess what? He's a bad guy! It's a miracle!

Not that that's what's happening here, but it's really similar to how it works now, only it's the sort of thing that would really be a great way to push that info to a LEO and the LEO never know it happened. "Ok, we're getting a hit on this guy with a mustache and briefcase in waiting line 3 - it says "this guy is acting odd" ok, go pick him up" and back at NSA central, some analyst checks off a box and goes to the next one.

PKD must be rolling in his grave.

i wonder what sort of knock-on effect this will have on social engineering
in the sense that people [as a group] tend to attempt to live up to their reputation [as a group]
....treat an area as a 'high crime' area and it will stay that way.
so they have 'statistics'
soon those will be 'established facts'
...because we totally needed more ways to divide people along arbitrary lines.

i like Getreadyalready's lottery analogy.
the mechanics behind it will stay just as occluded to the rubes
and the 'winners' will be announced on television, too.

Originally posted by jonnywhite
reply to post by Jerisa

 

I think it's worse to setup fake prostitutes on the street to nab lonely men. That's the worst kind of predictive crime because the crime might never have happened if it wasn't staged. I mean, the people doing the undercover work not only get money from the work they do but they also have a burning desire to arrest bad guys. In my humble opinion, it's out of control. Whether they setup fake dope dealers or whatever, it's the same thing to me. My opinion is that until a crime actually happens or is in its early stages and is not incited by authorities, any prosecution of a crime is suspect. We should instead focus on getting prostitutes off streets and to figure out how to discourage men. Do this without playing god. Without actors. Let the REAL crimes unearth from the chaos out there.

Now back onto the topic of hte program they use...

The program they mention is just giving them a heads up about what to expect from the numbers. It reminds me of an article I read about terrorist incidences and how they occur in special patterns. It was the July/August 2010 issue of Discover. The article follows Neil Johnson, a University of Miami physicist. It says he studies complexity. That doesn't mean much on its own. A lot of things are complex. But this kind of complexity has to do with how things mathematically behave in nature. So he's one - of a number of others - who studies various patterns in nature and utilizes graphs and models and math to explain them. The article also names people like him Quantitative Analysts. Neil's research into explaining warfare-type events is focused on in the article.

They discovered that terrorist events follow a power law curve and act like the rise and fall of the stock market. There're other events in nature that follow a similar pattern: weather events, income distribution, power outages, earthquakes, etc.The lack of a central command is a big part of the reason terrorist and insurgent groups behave in this manner. So what does this tell military strategic planners? It tells them that they cannot approach terrorism the same way that they approach conventional warfare against an enemy with a central command.

These kinds of things cannot tell you who will specifically commit a crime, but they can help you to have the right set of expectations and to approach a problem with better insight of its nature.

edit on 9-7-2012 by jonnywhite because: (no reason given)

Coercion is already illegal. It is also against the law for persecutors to offer sentence reductions for testimony or states' evidence, after someone has been accused of a crime..

These are on the books, but they continue to be SOPs, because officials are always seeking fast resolutions instead of justice.

in its pure form, there is nothing to indicate that there is a problem with studying trends and seeing where the data goes. I have no problem with using police in a predictive manner in order to prevent crime.

My problem is along that slippery slope that this seems to present. If you just follow the data, then that's fine. When armchair analysts decide they 'know' where the next trend develops, based on their prejudices and subjective analysis, then this runs right into the idea of using police as a offensive weapon instead of a deterrence.

Originally posted by MagoSA
in its pure form, there is nothing to indicate that there is a problem with studying trends and seeing where the data goes. I have no problem with using police in a predictive manner in order to prevent crime.

My problem is along that slippery slope that this seems to present. If you just follow the data, then that's fine. When armchair analysts decide they 'know' where the next trend develops, based on their prejudices and subjective analysis, then this runs right into the idea of using police as a offensive weapon instead of a deterrence.

Can you offer a plausible scenario?

Originally posted by g2v12

Originally posted by MagoSA
in its pure form, there is nothing to indicate that there is a problem with studying trends and seeing where the data goes. I have no problem with using police in a predictive manner in order to prevent crime.

My problem is along that slippery slope that this seems to present. If you just follow the data, then that's fine. When armchair analysts decide they 'know' where the next trend develops, based on their prejudices and subjective analysis, then this runs right into the idea of using police as a offensive weapon instead of a deterrence.

Can you offer a plausible scenario?

Lots of plausible scenarios. Police use statistics to know that more officers are needed in the bar districts just after last call. They use statistics to determine where traffic accidents are the most likely and place more traffic enforcment in those areas. They use statistics to zero in on patterns of crime and if they have a suspect list, they can cross-reference the known crimes and known suspects and see if anything stands out.

There is never such a thing as "too much information," but the issue is how to use it effectively without violating anyone's rights.

I wish they would say - they stopped me from doing a crime - talk about one hell of a law suit. where is their evidence they produce a computer program that says you are going to do something illegal,. and they you open you computer and type "And you will now desposit 20 million dollars into XXXXX bank account". if one is true then the other is true its all in the math,

The motto should really change to "Deny Reality". They now just have a program that maps out the high crime areas, the police have been doing this for years on their own. They are not going out to arrest people for pre-crimes.

Originally posted by getreadyalready

Originally posted by g2v12

Originally posted by MagoSA
in its pure form, there is nothing to indicate that there is a problem with studying trends and seeing where the data goes. I have no problem with using police in a predictive manner in order to prevent crime.

My problem is along that slippery slope that this seems to present. If you just follow the data, then that's fine. When armchair analysts decide they 'know' where the next trend develops, based on their prejudices and subjective analysis, then this runs right into the idea of using police as a offensive weapon instead of a deterrence.

Can you offer a plausible scenario?

Lots of plausible scenarios. Police use statistics to know that more officers are needed in the bar districts just after last call. They use statistics to determine where traffic accidents are the most likely and place more traffic enforcment in those areas. They use statistics to zero in on patterns of crime and if they have a suspect list, they can cross-reference the known crimes and known suspects and see if anything stands out.

There is never such a thing as "too much information," but the issue is how to use it effectively without violating anyone's rights.

What if the profiles of demonstrators were entered in an algorythm program after having been arrested and then placed under phone surveilance prior to upcomming political events in which they might again protest?

The title to this thread is a bit misleading. The police are the only people who use predictive analytics. Financial firms, social networking sites and advertising agencies all use predictive analytics to predict behavior.

en.wikipedia.org...