Schneier on Security
A blog covering security and security technology.
« MI5 Terror Alerts by E-mail |
| Choosing Secure Passwords »
January 10, 2007
Surveillance Cameras Catch a Cold-Blooded Killer
I'm in the middle of writing a long essay on the psychology of security. One of the things I'm writing about is the "availability heuristic," which basically says that the human brain tends to assess the frequency of a class of events based on how easy it is to bring an instance of that class to mind. It explains why people tend to be afraid of the risks that are discussed in the media, or why people are afraid to fly but not afraid to drive.
One of the effects of this heuristic is that people are more persuaded by a vivid example than they are by statistics. The latter might be more useful, but the former is easier to remember.
That's the context in which I want you to think about this very gripping story about a cold-blooded killer caught by city-wide surveillance cameras.
Federal agents showed Peterman the recordings from that morning. One camera captured McDermott, 48, getting off the bus. A man wearing a light jacket and dark pants got off the same bus, and followed a few steps behind her.
Another camera caught them as they rounded the corner. McDermott didn't seem to notice the man following her. Halfway down the block, the man suddenly raised his arm and shot her once in the back of the head.
"I've seen shootings incidents on video before," Peterman said, "but the suddenness, and that he did it for no reason at all, was really scary."
I can write essay after essay about the inefficacy of security cameras. I can talk about trade-offs, and the better ways to spend the money. I can cite statistics and experts and whatever I want. But -- used correctly -- stories like this one will do more to move public opinion than anything I can do.
Posted on January 10, 2007 at 11:36 AM
• 64 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
It's been mentioned here before, but security cameras can be used to create a denial of service attack on the police.
If the police see what appears to be an attack, they can't not respond and I guess that the US is little different to the UK and over night there are normally very few coppers available on patrol. Personal safety trumps property safety every time.
> I can write essay after essay about the inefficacy of security cameras.
Except that a single real-life story that indicates efficacy makes such an essay require quite some qualification. "It never works!" isn't true.
> I can talk about trade-offs [...]
Indeed, that's a much better (though complicated) argument.
>people are more persuaded by a vivid example than they are by statistics
But "bad guys" are people too.
I don't know that it is an actual DoS, though I guess those wishing to disrupt city services via large-scale civil disobedience could do such a thing.
However, it would be very easy to have an associate create a diversion (or 12) that the police would have to respond to, thus leaving the *real* target unprotected.
My boss is former FBI, and if you've read anything about the FBI recently (their pathetic computer system, lack of proper governance, poor and uninformed leadership) you'll know that they're not as on top of things as Hollywood likes to portray.
She said, the single biggest factor in successful crime fighting is the fact that most criminals are really stupid. That's it. It's not like they catch these sorts of masterminds like Frank Abagnale all the time. If all of them were that clever we'd be in real trouble. The good part is, criminals buy the same silliness that regular people do. The bad news: citizen's rights are being eroded pointlessly and result in less safety than before, because corrupt law enforcement et al abuse it.
Police cite surveillance camera as crime preventing device.
The cameras did nothing to prevent the crime... even if there was someone actually watching the live feed.
So no we are not safer with surveillance camera.
> we are not safer with surveillance
I guess the surveillance didn't help McDermott, did it?
But it did help law enforcement. If you're trying to convince the public that mass surveillance is a poor strategy, this law enforcement win makes for an inconvenient truth.
Constant mass surveillance is a powerful tool that can be used for good or evil, but mostly for ineptitude. Government hopes to replace expensive detectives with cheap cameras. Sometimes they'll catch the bad guy, but overall things will be worse.
The surveillance camera does nothing to protect against the crime observed, but it may provide evidence to use against the perpetrator. It's another tool for forensics people to figure out what happened and who did it. In this case it lead to the murderer, who was a loony.
On the flip side, a police officer watching from the same viewpoint would not be able to stop the crime either. From a rooftop, he wouldn't be able to give chase effectively without losing sight of the suspect. The officer could not effectively stop the suspect without a rifle, and even then his only option would be lethal force without knowing what happened.
Bad things happen. Surveillance can not stop that. Anyone who claims otherwise is selling surveillance hardware or services.
They don't seem to have caught this guy. It may help them that he seemed to be a contract executioner. So they can look for someone with a personal motive, not a random mugging or rape gone wrong.
Husband, coworker, high school enemy, mother in law, husband's girlfriend, whatever.
Personally, I'm not against the use of surveillance, just its ABUSE. In this case, the surveillance appeared to be used correctly.
If the FBI were to make the camera information available to the GOP so they could could monitor and collect dirt on Democrats, then we have a problem. obviously abuse of surveillance.
In my view, mass surveillance is not totally bad. It is the centralization of that surveillance, who controls it, and for what genuine and abusive purposes it is used for that is the issue, IMHO.
How is this effective? We have a dead woman and an assailant still at large. Sure, they have more data on which to pursue someone. For the time being, they have prevented nothing. Just like locks, cameras only keep honest people honest.
@Frank Ch. Eigler: There is a huge difference between "inefficacy" and "It never works". CCTV does work if integrated in a system with other prevention and response measures. A 2005 Home Office Research study in the UK on an number of CCTV schemes concluded that these had 'had little overall effect on crime levels'.
Surveillance cameras can also inadvertently cause harm. I remember reading about an incident many years ago about a guy (in London) that had a bad year. It was Christmas and his girlfriend had just dumped him and he had just lost his job. He was on a bridge and was going to kill himself with a knife. A camera caught him and the police responded and his life was saved. However, the camera footage got to the media and the guy's life was ruined. Everywhere he went to get a job he was turned down because he was the "guy that tried to kill himself". Granted, the cameras themselves didn't cause this issue, but the lack of security on the footage (camera footage was never supposed to be released to the public) was inadequate. I can see this happening more often than not in the US.
Wasn't it baltimore whose crime rate drastically fell after they installed cameras on the street corners?
You will never stop crime, but the more often people here others getting caught, the more someone will have to think about whether the risk is worth it.
The full story states that they did catch the killer, and that he had killed several others in the same way--seemingly random acts of violence against (in all but one case) strangers.
The logic of omnipresent surveillance is that it 1) serves as a deterrent and 2) serves as an investigatory tool. Obviously, here 1 failed and 2 worked, as the killer either didn't know or didn't care about the surveillance (based on the story, probably the latter).
In this case, surveillance proved its potential value, but I still can't shake the ease in which it can be used to control society. The reporter also stated that Americans are becoming used to the ubiquity of surveillance, and that seems to me a bad thing. We need to be able to choose to have our privacy, not to have it taken away from us little by little.
> a police officer ... would not be able
> to stop the crime either
I think you are selling this hypothetical police officer short. Why you have him sitting on a roof I don't understand, but if you were a cop and observed some guy stalking a lady down a dark and empty street, would you just sit there feeling impotent?
@ Straw Man
I haven't watched the video, but from the description a cop on the street wouldn't have been able to do much here either.
Even if said cop had thought the guy's behavior was suspicious (and there's nothing here to indicate that it was compellingly so), he'd have to follow the suspect and victim around the corner and remain within a pretty close distance to effectively attack and disarm someone who suddenly whips out a firearm.
From the description, the attacker acted with unexpected suddenness. It's probably doubtful that a cop could halt that crime in process*
* unless his presence was enough to dissuade the attacker, but if so the attacker would most likely have just broken off following the woman and chosen another victim later while there was no officer present, just causing a delay, not prevention.
This is an interesting feature of human nature that interferes with rational discourse. Just bringing it into wider awareness helps to elevate the discussion, so I hope you can find a more catchy meme than "availability heuristic".
You're saying that recent (or often repeated) (virtual/media) exposure magnifies perceived threats out of proportion with actual risks.
I'm hunting for a pithy phrase, the way "security through obscurity" communicates instantly. Can't quite blend in the role of media:
"Fear is an inaccurate measure of risk."
"The 'as seen on TV' threat barks louder than its bite."
"The oft-reminded threat looms larger than its real-life risk."
"A threat to mind grows the fear, in time."
OK, so I'm no Poor Richard. Anyone else wanna try?
> the single biggest factor in successful crime fighting is the fact that most criminals are really stupid
Absolutely! I worked for a major police department for nine years (network admin and database developer/admin) and BS'd with officers at every given opportunity. Retirement parties were especially good because they always told great stories. It's frightening to think how many smart criminals are probably roaming the streets.
>to effectively attack and disarm someone who suddenly whips out a firearm.
It doesn't work that way. If someone has a gun out and is threatening someone, an officer is totally within his rights to shoot them. Their job is to defend themselves and the public against the illegal use of lethal force. There is no such thing as disarming someone with a gun, it cannot be done without endangering the officer or civilians.
Basic scenario: you are an officer, your gun is in your holster. About 15' away from you is a nutcase with a butcher's knife. If that guy attacks you, you will not get your gun out of its holster before you're stabbed/slashed, and kevlar vests offer minimal to nill protection against knives. The recent upsurge in police departments equipping everyone with tasers is largely as a response to this situation. One officer can talk to the subject and try to calm him down and get him to drop the knife while others get their tasers out.
Law enforcement, sadly, doesn't do much to prevent crime. When you call 911, it's during or after a crime, not before (well, usually). A police cruiser going down the street might dissuade someone from a mugging or burglary, but if the officers are not visible, there's little preventative effect. This is where neighborhood groups and taking responsibility for your own (and your household's) defense come into play.
> If someone has a gun out and is threatening someone, an
> officer is totally within his rights to shoot them.
Oh sure, but I doubt your average cop would be suspicious enough of Joe Stalker here to have enough time to draw, aim, and shoot Joe before Joe shot his victim. I did say "attack and disarm", which would include "shoot and kill" :)
I expect that within the next five years we will see "America's Most Disturbing Survalence Videos".
Everything you say or do with become fodder for the voyeristic public.
This is an interesting case because it was solved not by the sexy part on which the reporting focuses: piecing together over a dozen private surveillance tapes across the city as they traced his escape. This has nothing to do with the arrest because it quickly turned into a cold lead. ABC News glosses over that logical gap.
The bus driver watched the tape during a police interview and thought she might recognize the killer. Police now had a suspect and discovered that he worked at the same facility as the victim. Police then reviewed surveillance video of the workplace and saw him walking into work in the same attire about thirty minutes after the killing. During interrogation, the suspect confessed.
What do we have here? The good old police work that Bruce and others routinely recommend. Interviews of potential witnesses, cops working the beat interviewing locals and collecting eyewitness reports (private surveillance videos), suspect round-up, and effective interrogation.
The fact that the ‘composite sketch’ was really a private surveillance video is almost beside the point.
It should also be noted that only one of these videos was a police camera. The rest just archive data automatically without anyone watching the feeds. The data is only pulled if something happens. That doesn’t bother me as much as state-run city-wide surveillance. In fact, it probably has the capability to exonerate more people than it puts behind bars.
Businesses can refuse to offer the tapes unless ordered to do so by a court order. Police have to expend a lot of manpower to track down, assemble, and analyze tapes. This means they are less likely to abuse this permanent surveillance record because of the required manpower. It isn’t perfect but it isn’t Orwellian.
The London bus/train bombings, the failed Cologne train bombing: more anecdotal evidence of surveillance camera's used after the fact to detain suspects.
I would like to know how frequently surveillance video is used to catch suspects. How many are later convicted? How many are acquitted? How many would have been caught and convicted even without the footage? (that's harder to answer). How much money is spent on video surveillance per conviction? Have criminologists studies this?
> The fact that the ‘composite sketch’ was really a private surveillance video is almost beside the >point.
Composite sketches usually aren't that helpful. Videos can capture a person's overall appearance, including gait, posture, and habitual movements. From this we can often recognize a person we're familar with, even if we can't see a face.
I think there should be cheap, ubiquitous surveillence systems spread throughout high crime rate areas. However, their recordings should be available to law enforcement officials only, and only with a warrant.
For example, a detective could submit a request to a judge for all surveillance feeds from a 4 block area over a 5 hour time frame after providing probable cause of a specific crime committed. Typical search&seizure rules would apply: the detective could only use evidence in the camera to prosecute the crime used for the warrant.
Similarly, live feeds could be requested much as wire taps are used today: with probable cause that they will aid in investigating a particular crime.
Bruce, read the book "Stumbling on Happiness" by Daniel Gilbert. He's a psychologist, but it's a pretty witty book anyway (perhaps a little too witty in places).
He mentions "presentism" as a common human failing, where presentism is defined as the tendency for current experience to influence one's views of the past and future.
I'm about halfway through tbe book, and I hadn't been thinking of what he's written in security terms, but there's a lot of it that is applicable to security. One of his main points is that these flaws are intrinsic to human thought, in part because our brains are "imagination machines" (my term, not his) and even what we perceive as "recollection" is actually key points recollected and embellished by imagination to "fill in" the rest. Indeed, our real-time perceptions also have this "filling in", something I've also read in Daniel Dennett's "Consciousness Explained" (also an interesting read).
Have you checked out any popular TV channels recently? Popular press? Popular booklists?
Bruce, you raise a good point. But it also suggests a counter-strategy. If you want to advocate for privacy, talking about tradeoffs is likely to reach the audience you have now, people already interested in security. If you want to persuade the public at large, vivid, real-world examples of your own are likely to be more effective.
Ubiquitious surveillance cameras don't bother me...as long as the data collected has short retention periods.
Someone holds up a convience store, go to the tape.
The problem is in this day of DVR, many of the units I see advertised retain the data for nine months or more.
Someday, maybe 10 or 15 years from now, that's no longer going to be stuff that's only looked at if there's a crime. It's going to be a commodity private businesses can sell to data aggregators.
Sure, the first uses might be fairly anonymous -- image analysis software to determine what the teens are looking at in the aisles, what marketing displays are effective. Over time, the analysis gets better and no "anonymous" images are being correlated from store to store. Credit card records reveal some of that, but it doesn't reveal what you're looking at, or which aisles are being visisted.
Ok, tin foilish...maybe.
But the algorithims one day WILL be powerful enough we don't have to worry about wearing RFID equipped sneakers...image analysis will track people by clothing and body type as they go about their shopping.
Privacy protections should be in place that specify how long it's proper to retain the data, and that it's only to be used for timely security purposes.
"I'm in the middle of writing a long essay on the psychology of security."
I think I've mentioned this before but I seem to keep getting asked to work on encryption projects rather than the human element of security...maybe we're all destined for a happy medium.
"...the single biggest factor in successful crime fighting is the fact that most criminals are really stupid"
Yes, or you might even say all criminals are stupid because they are convicted. If they are not convicted, then you can't really call them criminals, can you? Or maybe that's just a tautology.
This also makes me think about Bruce's paper in a reverse light. The more experienced an attacker becomes, the more proficient they probably become with the risks. So this would suggest there should be opportunities to make attackers more "stupid" by persuading them with vivid examples.
"Ubiquitious surveillance cameras don't bother me...as long as the data collected has short retention periods.
Someone holds up a convience store, go to the tape."
So, you saying you are in favor of quick convictions? What if there exists some doubt about the surveillance tape and time is needed for further research/review/correlation? Do you then allow for someone to preserve the collected data while more data is collected or do you go right to a verdict in order to stay within your definition of "timely"?
@ Aaron Muderick
Yes, good points. However, I'm surprised you didn't conclude that the "good old police work" you mention was a process of collecting available information including private surveillance camera data. Seems to me that justifies it, although I agree it should not be seen as sexy any more than traditional witness accounts and/or informants should be...
"Fear is an inaccurate measure of risk."
Don't be so hard on yourself. I thought that was quite good - simple and to the point.
"I think there should be cheap, ubiquitous surveillence systems spread throughout high crime rate areas. However, their recordings should be available to law enforcement officials only, and only with a warrant."
Interesting point of view but I see a potential clash between economics and privacy developing. If the surveillance system is to be cheap, then there will be a temptation to skimp on security to get market share e.g. I am a bar owner who wants cameras and has to choose between professionally installed and secured system versus the bargain Wal-Mart offering. It's not too difficult to imagine a situation in which criminals might hack into these systems because of their cheap and nasty security. Which naturally leads onto the next comment...
"lack of security on the footage ..."
I think this is a good point; If we could trust all CCTV operators then there wouldn't be a problem with mass (private) surveillance. Perhaps you should be licensed or something before you are allowed to use CCTV technology in public places?
Thomas Kida wrote a book called "Don't believe everything you think: The Six Basic Mistakes We Make In Thinking". I haven't read it, and to save you the trouble, here are the six basic mistakes:
We prefer stories to statistics; we seek to confirm, not to question, our ideas; we rarely appreciate the role of chance and coincidence in shaping events; we sometimes misperceive the world around us; we tend to oversimplify our thinking; our memories are often inaccurate.
Sound familiar? Seems like we spend a lot of time picking apart faulty thinking in this line of work, regardless of the rightness or efficacy of CCTV surveillance (and many of the specific issues you point to daily).
> Do you then allow for someone to
>preserve the collected data while more
>data is collected or do you go right to
>a verdict in order to stay within your
>definition of "timely"?
Short retention of data isn't short retention of evidence.
Once data is identified as actual or potential evidence for a specific act, it falls under evidentiary rules and can not be lawfully destroyed until it's evidentiary value is determined.
Hardly a success story.
They had to consult dozens of cameras across the city, and even then they only caught the crimminal because of a tip by the bus driver.
If anything, this is proof that whole-sale surveillance via cameras don't work. You need a team of police officers, dozens of cameras, and you still can't identify the crimminal.
Your tax dollars at work.
>Police cite surveillance camera as crime preventing device. The cameras did nothing to prevent the crime... even
>if there was someone actually watching the live feed. So no we are not safer with surveillance camera.
Are we reading a different story here? According to TFA, the guy was a serial killer, and was caught this time only because there were two differenct surveillance cameras (the one at the scene of the crime and the one at her workplace) that provided evidence necessary to catch him.
It didn't prevent *this* crime but it caught a serial killer who would have very likely killed again. So, it has a measurable positive effect: serial killer put behind bars due to critical evidence collected by surveillance cameras.
The more appropriate question is: is this positive effect worth the trade off of monetary costs of cameras, and societal costs of always being spied upon by anonymous watchers?
Bruce is right; I think in the right hands this is a boon to the pro-surveillance folks. Lots of people would rather be stalked by anyone who can get their hands on a security tape, than keep their privacy at the risk of a tiny chance of being shot in the back of the head by a nut.
Are there any colorful stories about security camera stalkers abusing their information to assault/rape/murder someone? That would be a suitable counterbalance to this story.
I'm of mixed feelings on this issue. But I am certain that most people haven't thought it through. It's good to see this discussion.
People get very involved and emotional in discussions of organized surveillance.
I am also concerned about the potential for abuse and in favor of short retentions and strong controls. But the basic questions are interesting and the discussions valuable.
It seems to me that we already live in a world where increasingly we are under private surveillance. Whether CCTV, cell phone cameras, or small CCDs. The technology is getting smaller and more pervasive all the time.
Many of the people taking these images already abuse our sensibility of privacy. Anyone in a public square on New Year's Eve really should have no expectation of privacy and has no justification to claim their privacy was invaded. Now switch context to a company Christmas party or college party. Consider how this changes?
In Toronto, this last month, there were CCTV's installed downtown as an experiment. The cameras covered public spaces. Partly this was a reaction to the fatal shooting of a bystander, Jane Creba, a year ago. It took over 9 months and a huge effort but the cops caught the shooter. This year these cameras caught another gun (or knife) attack just a few yards from the last shooting. Public opinion has ranged from "in full support" to "invasions of privacy". The cameras are now gone.
I agree that CCTV's will not prevent attacks. They may perhaps deter them. Maybe. The biggest benefit to law enforcement would appear be the potential to solve more cases and to solve them at less cost to the tax payer. Perhaps it may help prevent repeat offenders.
My question is how real is this? Will they solve more crimes? At less expense? When all is considered? And, what other problems are we buying into?
These questions are both harder and deeper than a cost benefit analysis or the argument that it is worth it to catch a killer that would otherwise get away. They need to be examined carefully.
@Unonymous -- Get with the program! The Schneiermeister already recommended "Stumbling on Happiness" last December.
"Yes, or you might even say all criminals are stupid because they are convicted. If they are not convicted, then you can't really call them criminals, can you? Or maybe that's just a tautology."
Of course you can call them criminals if they're not convicted - you know a crime was committed (for example, a car was stolen), you know that someone must have stolen it - while you don't know specifically *who* the criminals are, you know that the criminals must exist.
Remember also that the concept is "presumed innocent until proven guilty" not "innocent until proven guilty" - the accused can be guilty, but the legal system requires that we prove it. But that doesn't mean that I as a non-juror can't refer to the accused as a criminal before the jury makes its decision after a sniper holed up in a building is caught when the cops rush the place.
@ wkwillis, TorUser
They did catch him using a combination of piecing together various surveillance cameras and a tip from a bus driver familiar with the murderer. It turned out he was a serial killer nutcase, not a hitman.
Interesting what can be learned from the report. For example, changing your clothes/appearance to make it harder to connect the data from diverse cameras is pretty important. Of course this isn't anything new, look at David Brin's novel "Kiln People", for example, for his extrapolation on the kind of things people in the future may do to "lose" this kind of surveillance (not the best example, I suppose, because of other unique elements he added to that future).
That London Metro poster is ... Orwellian. Fritz Lang would be proud of that imagery.
"Are there any colorful stories about security camera stalkers abusing their information to assault/rape/murder someone? That would be a suitable counterbalance to this story."
I'm not so sure that it would. People abusing access to the tapes is a systemic/procedural issue, that people can and do rationalise quite easily.
Also, abuse of a tape isn't quite the same as an actual murder on the street. People will weigh up their personal cost/benefit - i.e., chance/cost of some loon might get their rocks off watching a tape of me walking down the street vs. chance/cost of some loon shooting me in the back of the head. The chance of the first might be high, but the cost is low. Chance of the second might be low, but the cost is rather high. Reasonably naturally, and apparently logically, a lot of individuals will decide that - for them - mass surveillance makes sense.
Of course, that analysis breaks down when you look at entire populations, since the cost becomes "everyone is now an unwitting porn star in waiting" and the benefit is "the police might catch a bad person. After they've been bad. Maybe. If they're lucky, and do a bunch of other work - investigation, following leads, etc. - that they would have to do anyway, with or without the mass surveillance."
Is this a tragedy of the commons situation? Mmm. Probably not, but it's kinda similar.
There is an extremely simple argument to be made here:-
"There a 100,000s of surveillence cameras deployed througout the world; yet a survillence camera doing what it was designed to do still makes headline news."
> It's frightening to think how many smart
> criminals are probably roaming the streets.
The smart ones are not roaming the streets. They are elected to the high offices.
Everybody is talking about the inefficiency of surveillance systems. I agree, surveillance cameras don't prevent random acts of senseless violence or terrorist attacks. But that's not their purpose. Most surveillance cameras monitor private property, factories, warehouses, shopping malls and supermarkets etc. Terrorists aren't stopped, but burglars, shoplifters and intruders are caught.
Surveillance systems are a valueable asset to the property owners - but they have nothing to do with increasing public safety. The problem is that the general public is lead to believe otherwise, that the streets will somehow be safer if there are more cameras.
Another thing to point out is that while people are freaked out about Big Brother and mass surveillance, the real big thing nowadays is Little Brother - the citizens themselves are now recording each other's every move with mobile phones and other devices. Whenever something disturbing happens on the streets, it's the mobile phone pics and videos that get uploaded to millions of blogs and eventually make it to the front pages.
@ David (Toronto) wrote:
> I agree that CCTV's will not prevent
> attacks. They may perhaps deter
> them. Maybe.
I'm of the firm opinion that it's the probability of being caught that deters. If people (note I said people, not criminals) think that cameras increase their chances of being caught doing something they shouldn't (speeding for example) then they are less likely to do it - at least where they suspect there might be a camera.
However, many businesses have security cameras and still get robbed.
The footage shown on TV from the cameras often seem to be of poor quality so I'm not sure that it leads to that many convictions. If the images are poor quality, not only might it be difficult to identify the individual, but it might also lead to false positives.
Even if the pictures are of very good quality, how do you put a name to the face?
I think the footage is most useful if the Police already have a suspect, or if it contains addition information that can identify an individual (a license plate for example).
As to the privacy of the footage, I suspect Police PR departments like giving the footage to the media to gain favor with them.
To those saying the cameras don't prevent crimes -- you don't know that. They didn't prevent *this* crime, but there's no way of knowing how many potential criminals have thought twice knowing there are cameras about. I'm not arguing for cameras, I'm just pointing out that one cannot assert that they do nothing.
I'd also point out -- and this is in support of Bruce's reason for posting this -- the camera's involvement in the story is pretty minor. It was a pretty routine investigation that happened to be resolved more quickly by his being caught on camera. Giving the headline to the camera was a questionable call
On the flip side again, if you read the whole article, it turned out this person was responsible for *many* murders, not just this one. Only when caught on camera was he captured, possibly preventing him from killing more.
"It explains why people tend to be afraid of the risks that are discussed in the media, or why people are afraid to fly but not afraid to drive."
Bruce, I disagree with the premise. Considering the sheer number of traffic accidents, I think nearly everyone in America knows someone personally that has been injured, maimed, or killed in an accident. I happen to know quite a few, and even a few who were horribly disfigured or maimed. Also, most of us have witnessed an accident or two in person or been involved in at least a fender-bender.
Compare that to 9/11. ~3000 people were killed in one tragic day. Outside NYC most people know of people who knew people that were killed, but very few in comparison were personally affected.
The main difference between car accidents and crime/terrorism, is that most people FEEL that they are in control when driving a car. When confronted by the dire consequences of death by murder or terrorism (or hunger after a hurricane floods the city), people over react because they don't perceive themselves to be in control in that situation. To get around this feeling of insecurity, people with the imagination to see themselves in a similar situation call on the nanny-state to keep them safe and (if they haven't become one of the Living Impaired in the process) are surprised and outraged when it doesn't.
"Once data is identified as actual or potential evidence for a specific act, it falls under evidentiary rules and can not be lawfully destroyed until it's evidentiary value is determined."
That's a pretty low bar, since potential evidence would cover a lot of data. Moreover, you've made another interesting turnabout as that pretty much undermines your point about short retention periods. In fact, with SOX and similar regulatory requirements spanning an entire year of activity if not longer (not to mention the rather long average length of time between events and investigations), you've just made the perfect case for long retention...easier to retain everything as potential evidence than destroy selectively and be accused of some particular intent, eh?
Fair points, but I don't think you've reached a conclusion that smart criminals exist, or that there can be such a distinction. That was my worry about the tautology.
"...while you don't know specifically *who* the criminals are, you know that the criminals must exist..."
Or in other words you know that there are unnamed attackers out there who will be caught and convicted, because they are criminals.
BTW, unless I'm missing something in the comments above, I haven't seen anyone suggest yet that the surveillance helped rule out innocent people. In other words, the bus driver, the witnesses, the police all had some pretty strong evidence that would eliminate a huge percentage of the population from consideration.
So unless someone can show how the integrity of the surveillance data in this case was compromised, the cameras perhaps help protected countless innocent people from the threat of false accusation.
This has often come up in data centers where I have expanded surveillance controls, as staff often appreciate that cameras serve as a form of witness and thus can help them demonstrate their innocence if necessary.
While it's true that a camera does very little to stop crime, on the other hand, a camera equipped with a remote-controlled high-power rifle or better yet, machine gun... hmm.... that would be really cool, wouldn't it? :)
In my opinion, all public spaces should be monitored continuously, the footage archived permanently, and available without restriction to anyone who cares to watch. We're heading in that direction anyway, so we might as well make sure it's done right.
How to prevent the extension of ubquitous surveillance (which is inevitable) to ubiquitous law enforcement (which would be a disaster) is an interesting question.
It should be SOP that each member of the military on combat duty (Iraq, for example) wear a small camera on their helmet at all times. It would help foresically with the "we werent shooting at you, we were just celebrating a wedding by firing belt-fed crew-served weapons, RPGs and 60mm mortars into the air at random" events.
All police vehicles and interrogation rooms should have cameras recording at all times to protect both sides in any follow-on legal proceedings.
Police speed guns should have an integral camera with a tamper-resistant clock that shows when the speed reading was taken and what vehicle (or signpost, airplane, UFO, whatever) was in the bulls-eye at the time.
Unless the cameras are everyone, they can only catch crimes committed in front of them. And do we really want a world in which there are cameras everywhere?
A camera is useless to identify someone if they simply where a hood or other disguise.
If you have seen a movie, you know that what you see on camera isn't necessarily true. The crime can be fake, or the perpetrator can use it to help frame another by dressing or otherwise acting like the person being framed.
In the end, the more government fails to trust its citizens, the less its citizens will behave nicely. The more government intrudes, the more citizens hide, trick or otherwise screw up the system.
This one example, like Bruce says, suggests cameras are good, yet there are hundreds of murders that are not captured on camera and likely wouldn't be as many (most?) murders don't happen in public places with lots of people, the very areas cameras would be located.
Amazingly, Timothy McVeigh was caught on camera, but they weren't government cameras. In the end, giving the government this mandate and power to watch is likely to result in abuses that are harder to control with lots of cameras owned by disjointed people.
Just as the TSA confiscated my peanut butter and jelly -- resulting in a security failure as these were not weapons -- every hour that the camera doesn't catch a crime is also a security failure, recording non-criminal activities.
And of course cameras are expensive to install, operate and maintain.
"Fair points, but I don't think you've reached a conclusion that smart criminals exist, or that there can be such a distinction. That was my worry about the tautology."
"Or in other words you know that there are unnamed attackers out there who will be caught and convicted, because they are criminals."
Actually, my point is that while the vast majority of criminals are dumb, there have been documented instances of smart criminals (most of the ones we hear about are deathbed confessions & the like).
So in some cases, we have crimes that we look at the available evidence and say "Either this criminal is one of the smart ones, or he got damn lucky." For the lucky ones, either their luck runs out, or they realize they were lucky and quit while they're ahead. The smart ones often realize that there is a bit of luck involved and go for one or a few big scores and quit.
The tradeoff between technology and staffing is complex for the public police. Visible surveillance cameras do influence some criminals, I know for I have interviewed some of these folks in my real job (senior police commander in a major city in the USA).
Are surveillance cameras a panacea for crime? Absolutely not. However, it is a useful tool in so many contexts as to ensure continued application (perimeter camera patrols of a closed college campus, parking lots, multi-story garages, certain public streets, etc).
Folks will continue to link such information to their opinion in such topics, for stories are more compelling, especially as contrasted to a table of statistics. Yet the value of stories as an information vehicle have been well known for centuries.
Surveillance technology will continue to be a difficult public policy decision, I think the solution may be found in soliciting more input from the public that we serve. But raw public input is often skewed and not reflective of the opinion of the community as a whole. Difficult choices in dangerous times.
Well thats my two sents, probably with a third penny thrown in for good measure.
I would explain it this way:
What if the camera happened to catch the person being shot but not the shooter? What if the shooter wore a disguise? Or, and this even more likely, what if you put up a camera and no one even wanted to shoot anyone else in that area?
It's easy to come up with examples of people who've won the lottery, and it's exciting to imagine winning it yourself. Don't let that fool you into buying a bunch of lottery tickets.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.