Slashdot Mirror
Chicago Developing 'Suspicious Behavior' Monitoring System
narramissic writes "Over the past few years, Chicago's Office of Emergency Management and Communications (OEMC) has been blanketing the city with a network of thousands of video cameras in an effort to remotely keep track of emergencies in real time. Now, with the help of IBM, the network is getting some smarts. IBM software will analyze the video and ultimately 'recognize suspicious behavior,' says OEMC spokesman Kevin Smith. 'The challenge is going to be teaching computers to recognize the suspicious behavior,' said Smith. 'Once this is done this will be a very impressive city in terms of public safety.'"
...welcome our Minority Reporting overlords.
You have to field test your research somewhere, this one just happens to have a big juicy contract with it probably.
Robot Officer: Dave, you are under arrest.
Dave: What for?
Robot Officer: Hal says you are acting suspicious.
Dave: Picking my nose is acting suspicious?
Robot Officer: Yes you might be about to litter.
This is not a good idea now the cops can just say the computer said you were suspicious so we have reason to detain and search you and your car.
Should get really interesting when they integrate this system with the latest US Army battlebots!
Chicagoans should go out of their way to act "suspicious" in front of these cameras if they want to prevent the onset of a nanny state. Wear thick coats during the summer months, keep their hands in their pockets, look back and forth. Hell, maybe sticking their tongues out at the cameras would constitute suspicious ...
Besides, where they ought to be placing these cameras is in the halls of Chicago's city government.
!#@%*)anks for hanging up the phone, dear.
Cool, so, we're not even pretending anymore that the use of cameras are anything less than the complete and total expansion of the panopticon, are we? I mean, of course, you'll still have the people who say "well, if you aren't doing anything wrong, why are you worried", but for the most part, we're pretty up front about the fact that we're going to be using cameras to keep our citizenry under the thumb. Who defines what constitutes "suspicious behavior", local cops, politicians, computer techs? There will be essentially zero guidelines for the implementation of this technology, so what's to stop the local PD, or the DEA from auto-flagging someone who looks like they're raising a pipe to their mouths, or, even better, engaging in nefarious acts like leaving the house late at night? And not just that, but how many citizens will have their rights violated because of a false positive from the "suspicious behavior" flag? Will the flag be enough to get a warrant to search someone's car or home?
End of the fucking universe, right here.
When information is power, privacy is freedom.
'Once this is done this will be a very impressive city in terms of public safety.'
It will also be impressively Orwellian and unnecessary. I'm waiting for those famous Midwestern militias to get determined and start systematically tracking and disabling these cameras so that the rest of us can continue to go about our business w/o the prying eyes of the government.
I'm tired of traffic cameras, red light cameras, and the government's position that you are in the public and thus not anonymous in your actions. That rhetoric worked when you were manning more human police officers to do the work, not when you decided to become lazy and act like the public are your DVR favorites for watching and scanning at a later time.
I'd wager the false positve rate is going to be very high, and it will be interesting to see if they can bring that down. Something like an alert for a stolen car ( or a car related to an amber alert ) could generate a very high false positive rate if the car is a common make/model.
On the other hand, if it teaches criminals to act in less "suspicious" ways, then the system will be of no value or perhaps even detremental ( showing no "suspicious" behavior when criminal activity is present, leading to a false sense of security ).
A Human Right
City camera is pointed at a window in which is visible the screen of a computer at a public internet cafe. You log on to /. at the cafe. Bam! Suspicious activity! See, it's applicable. :)
Also, and call me crazy if I'm crazy, but its awfully hard to live as a free, responsible person online if you can't live as a free, responsible person offline . Hence, meat-space rights are relevant to YRO.
All the techniques ever used to make men moral have been themselves thoroughly immoral... (Nietzsche)
Wasn't there an article on how the massive London camera network doesn't actually do any good? And that one has real people monitoring it. Who really thinks a computer will be able to do a better job at something so nebulous as "suspicious behavior?"
Oh, that's right, nobody. However, that doesn't stop the company pushing this from trying to make a buck. It's sorta like the DRM companies. The DRM companies all know it doesn't work, but companies keep falling for the salesmen's lies.
How many times have I heard this on an AI related project? "Once this is done..." is a fantasy, especially when they already describe it as a "trick" and a "challenge." From TFA:
Challenge, indeed. I'll believe it when I see it.
Scratch that. I'll believe it when the system sees it.
"'Once this is done this will be a very impressive city in terms of public safety.'""
Impressive if your main hope in life is to live in some sort of Orwellian nightmare. Hey, here's a thought. If you put cameras in every house you can cut down on child abuse! You don't object to that do you? What are you some sort of kid toucher? Won't somebody please think of the children!
So much for Chicago being the lovely city I wanted to visit again.
If you wanna get rich, you know that payback is a bitch
We can't even make a translator that works and you're going to look for "suspicious" people? Is that some kind of sick joke? What, exactly, constitutes "suspicious" and wouldn't that be almost impossible to code in any meaningful way? Hell, we humans can't even agree on what "suspicion" looks like and now they want to teach a computer. Good luck with that.
I expect that "suspicion" is a fairly complex process in the human brain (it relies on a lot of different senses) so I am having difficulty understanding how anyone in their right mind would undertake such an effort.
... and unless we've made astounding progress in the last 5 years (as in, someone created a strong AI), IBM is full of crap. Completely, utterly full of guano.
Here's how the system will work.
head covered: check
metal flash: check
loud sound: check
Result: sound warning
There's absolutely no way in hell that the system is going to be able to do a real-time analysis that goes beyond basic image and sound recognition that is coupled with a set of expert rules. Why? Because no will have the time to properly train the system. And even if someone would be insane enough to do that, it'll still fail, because context is utterly missing.
Example: someone runs out of a store in a hurry. Someone comes after him. Should the police be involved? Did someone steal something from the store, or did two people find out one of their friends is in trouble? Or are they late to a movie?
This system is doomed to complete failure and is nothing but a boondoggle for IBM. Kudos to the IBM salespeople who sold Chicago on this system. They're able to sell fridges to eskimos, I'm sure.
The only thing that really worries: the politicians who drank the kool-aid. For those of you who live in Chicago: vote them out, or move. This is a sign of serious trouble on the horizon.
Those who can, do. Those who can't, sue.
Cars parked where they aren't supposed to be... cars that drive around the same building several times... obviously none of these people have ever been to Chicago.
In Chicago driving around a building several times is what you do before you decied to park somewhere you aren't supposed to be parked.
Sounds like putting cameras in the forest looking for trees.
If the system for example could recognize signs of someone being followed, it might be enough to dispatch a police car to drive past or ask the person being followed if they want assistance to help avoid a lot of serious crimes from being committed.
Now, there's still room for abuse (train the system to recognize likely politically unpopular groups and send police to intimidate, for example), but that doesn't automatically mean that there can't be ways of making this system useful without making it intrusive.
As long as you and everyone else keep waiting, it will never happen. Change occurs when people get fed up and do something about it themselves rather than waiting on someone else to solve the problem for them. If we hadn't been so gung-ho as a nation on giving other people the responsibility to protect us from terrorists/criminals instead of having the balls to take care of ourselves, then perhaps we wouldn't be in the position of needing to hope that a different set of other people will take responsibility to get our rights back now that we gave them all away.
Here's the discussion about that article. Plenty of opinions on both sides of the issue there.
You are aware, of course, that few things are as suspicious as pointing out that the Emperor(s) strutting around naked, right?
The one thing I constantly keep hearing about is all those police officers who show up to work day after day with nothing to actually do. This system will help those cops fill up the massive gaps in their daily schedules...
It's a simple matter of complex programming.
Hmm. Wouldn't it be cool if the headline was actually "Chicago Developing 'Superstitious Behavior' Monitoring System"?
http://outcampaign.org/
This is just silly. I understand that people would prefer to PREVENT crime instead of REACTING to crime, but you can't PREDICT crime as an alternative. Prevention and prediction are two very different things.
To prevent crime, educate the populace so as help to instill acceptable ethics and a sense of shame. Help them to acquire the resources they actually need and stop telling them they're less of a person if they don't have the "best" of everything. Teach people about people and reinforce those teachings throughout life.
To predict crime, go see a psychic because they are just as likely to choose an imminent criminal due to "suspicious" activity. You'd spend less money this way. You'll need it for the counter-suits.
Truth of the matter is that the nation isn't a fan of raising their children. Nor do they look kindly on higher taxes to reduce classroom size so that teachers can be mentors as well as lectures. And since crime prediction is a fantasy, the best we can do is crime reaction.
You are awash in a sea of fiercely stated opinions. Obvious exits are: 'File->Quit', 'Reply', and 'Page Down'.
On tv...now that gives me an idea! What if we pipe all the survelliance video feeds into the local access tv channels, flipping every couple of minutes and setup a hotline for citizens who watch the channels to call in and report suspicious activity! No complicated AI computer development and it has citizen oversight!
Demented But Determined.
Just walk around goofy...try to set huge ranges of standards for normal/abnormal behavior...throw the stats right out of the window.
That or everyone come to town on "dress like a stick of dynamite day"....
Light travels faster than sound. This is why some people appear bright until you hear them speak.........
Why? The role of the police department is not to prevent or stop crimes that are in progress. If someone breaks into your house and you don't have a weapon, you are likely fucked. Unless you can convince the burgler to sit nicely in a chair until the police come before he does anything to you. The role of the police department is to clean up and investigate AFTER a crime has been committed.
While these cameras might give you a little faster response, they're still not much more useful than providing post-incident records for court cases.
And frankly . . . while I hate the whole fucking big brother aspect that our society is taking since 9/11 . . . I'm getting too old to give a shit. Just bring on the inevitable and get it over with.
Let's apply a little Bayesian reasoning, shall we?
Given that system X identifies your behavior as suspicious, what is the probability that you are a terrorist? This probability is written P(T|S). This is what we want to find.
Bayes' Rule: P(T|S) = P(S|T)*P(T)/P(S).
P(S|T) is the probability that the system will identify you as suspicious, given that you are a terrorist. You can call this the system's "accuracy." Let's be generous and say the accuracy is 99.99% = 0.9999.
P(T) is the probability that you are a terrorist. Let's say that this probability is one in a million: 0.000001.
P(S) is the probability that the system thinks you are suspicious. There are two sources of suspicion: true positives, and false positives. The true positives are given by P(S|T)*P(T). The false positives are given by P(S|~T)*P(~T).
Let's again, be generous, and say that the false positive rate P(S|~T) is only 0.1%, or 0.001.
P(~T) is just 1-P(T) = 0.999999.
So, let's substitute everything in:
P(T|S) = P(S|T)*P(T) / (P(S|T)*P(T)+P(S|~T)*P(~T)) = 0.9999*0.000001 / (0.9999*0.000001+0.001*0.999999)
What's that equal, everybody? 0.0009989 which is about 0.001, in other words 0.1%
What does it mean? Even with a system that has a true positive rate of 99.99% and a false positive rate of only 0.1%, the probability of a "suspicious person" actually being a terrorist is only 0.1%.
In other words, these systems are inherently useless in identifying terrorists. This is because terrorists are inherently RARE in the population. The massive accuracy of the test cannot make up for this fact.
'...if only "Jumping to a Conclusion" was an event in the Olympics.'
It looks like this person is trying to sell crack. Cancel or allow?
That's our life, the big wheel of shit. - The Fat Man, Blue Tango Salvage
...were also processed like cattle in the 1930's, thanks to Big Blue. I am saddened that they haven't changed all that much, assisting a totalitarian government in having an omnipresent peering eye. Read up on it. The past is being repeated.
So, what about the mentally ill and disabled people? Will they be harrassed till they leave town?
Excuse me, but please get off my Pennisetum Clandestinum, eh!
Considering Chicago's political history, they'd better not put up any cameras near city hall.
Have gnu, will travel.
An equivalent amount of funding put into community policing programs or Neighborhood Watches would likely be far more effective than a camera program could ever be. When citizens start paying attention and giving a shit about what happens in their neighborhoods, things change. Buying lots of cameras and cops only seems to grow bigger tax bills.
If you never make mistakes, it's probably because you're not doing anything.
Some mornings it's hardly worth chewing through the restraints to get out of bed.
Assume one trillion possible indicators to sift through: that's about ten events -- e-mails, phone calls, purchases, web surfings, whatever -- per person in the U.S. per day. Also assume that 10 of them are actually terrorists plotting.
This unrealistically-accurate system will generate one billion false alarms for every real terrorist plot it uncovers. Every day of every year, the police will have to investigate 27 million potential plots in order to find the one real terrorist plot per month. Raise that false-positive accuracy to an absurd 99.9999% and you're still chasing 2,750 false alarms per day -- but that will inevitably raise your false negatives, and you're going to miss some of those ten real plots. source: http://www.schneier.com/blog/archives/2006/03/data_mining_for.html
Just so you know, many of the cameras are encased in a bulletproof box and have acoustic sensors to detect gunshots; the camera will automatically focus on the source of the shot.
The things transmit their video wirelessly and allow for remote control via a wireless link to police cruisers in addition to their wired link to the monitoring center downtown. I'd focus on jamming the signal or disabling the wired link.