Schneier on Security
A blog covering security and security technology.
« No-Fly List to Be Scrubbed |
| Friday Squid Blogging: Giant Squid Attaches itself to Yacht »
January 19, 2007
Sending Photos to 911 Operators
On Wednesday, Mayor Bloomberg announced that New York will be the first city with 911 call centers able to receive images and videos from cell phones and computers. If you witness a crime, you can not only call in -- you can send in a picture or video as well.
This is a great idea that can make us all safer. Often the biggest problem a 911 operator has is getting enough good information from the caller. Sometimes the caller is emotionally distraught. Sometimes there's confusion and background noise. Sometimes there's a language barrier. Giving callers the opportunity to use all the communications tools at their disposal will help operators dispatch the right help faster.
Still Images and videos can also help identify and prosecute criminals. Memories are notoriously inaccurate. Photos aren't perfect, but they provide a different sort of evidence -- one that, with the right safeguards, can be used in court.
The worry is that New York will become a city of amateur sleuths and snitches, turning each other in to settle personal scores or because of cultural misunderstandings. But the 911 service has long avoided such hazards. Falsely reporting a crime is itself a serious crime, which discourages people from using 911 for anything other than a true emergency.
Since 1968, the 911 system has evolved smartly with the times. Calls are now automatically recorded. Callers are now automatically located by phone number or cell phone location.
Bloomberg's plan is the next logical evolution -- one that all of us should welcome. Smile, suspected criminals: you're on candid camphone.
This essay appeared today in The New York Daily News.
Another blog comments.
Posted on January 19, 2007 at 12:22 PM
• 38 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
And here I thought you were always telling us that security cameras are bad security :)
Seriously, this seems like a really good idea and a sensible evolution of the 911 service.
A hand-held camera being used by a person who thinks he's seeing a crime isn't what the broad term "security camera" usually means! The individual camera doesn't produce tons of information that can be trolled (perhaps semi-automatically) later for anything incriminating.
Much of this has been developed at the University of South Florida under a few engineering institutes. They've also outlined a few other features, such as being able to warn people of natural disasters and Amber Alerts.
You have 100K (for example) people in the city carrying cameras and the city is putting up cameras and paying people to watch them. It might be cheaper just to give everybody a new camera phone and pay them to use it.
One small step for law enforcement, one giant step closer to wearing a "life recorder".
They could just let people upload pics live to a NYC website and monitor the site instead of having people send them photos. Then it's open source and more people can watch out. You wouldn't need operators involved.
A realtime site would be dynamic. One crime or other situation could draw hundreds or thousands of pictures in a few minutes. Realtime would let the police have them without the operator in the middle. It would be instant intel.
I hope the "futz factor" doesn't put anyone in danger -- such that the bad guy has additional time/opportunity to attack while the victim is trying to get a good snap to go to 911....
I hope the "futz factor" doesn't put anyone in danger
I doubt you do.
One slip, under the stress of getting the 911 vid in quickly, you accidentally send in a personal vid moment with your (girl friend, wife, mistress, spousal unit).
The safety bonus of this is amazing, too. A very difficult part of 911-operators' jobs is trying to ascertain the nature of injuries, severity of car accidents, and other such things which require long, drawn-out verbal explanations, but are easily captured with a single snapshot.
The ability to voluntarily share information is, once again, wholeheartedly welcomed. :)
@James: "It might be cheaper just to give everybody a new camera phone and pay them to use it."
Intriguing idea. You'd get better, more-relevant pictures (rather than the guy's trailing foot as he goes around the corner that the automated camera has just scanned towards in its mindless pan-back-and-forth sweep). You'd also automatically get the phone-number (and thus, name and address) of witnesses.
What are the privacy safeguards on the images supplied?
"innocent until proven guilty" doesn't mean much if you're convicted by a jury of you-tubers.
@James: Never thought anyone could out-cynic me, but you did it.
Might this be a new conduit for a virus or worm into the 911 computers? What kind of filters will be in place at the receiver?
Call me paranoid, but don't trust attachments.
The most useful part of this functionality isn't as evidence in a trial... camera phones are unlikely to provide much in the way of that, at least at this point in time.
But say I'm a little shellshocked after a serious car accident and I don't notice that the car I just took a picture of is smoking. The operator can see that in the picture, even if I forgot to mention it.
I think you're right that abuse won't be a big problem with the 911 service. I'm a bit concerned that it will be with the 311 version, however. (OMG, I actually read the speech transcript!) 311 calls can be somewhat terrifyingly powerful. At an old job, a building next door was belching foul-smelling smoke at our office; a good chunk of my co-workers actually went home sick because of it. I called 311, and within an hour a few Department of Environmental Protection trucks pulled up. The smoke stopped, the business was fined, and the guy who was improperly operating the boiler was fired. I was a little startled by all of this, but I don't think it was necessarily an incorrect outcome. I had a legitimate complaint, I think (hell, tens of thousands of dollars woth of lost productivity, at least), but in a city of what, 8 million, there are a lot of busybodies.
An area where this is quite apparent is in construction. People are always calling in to report illegal construction. Sometimes it really is illegal construction, and the photos would help that (often the inspectors arrive after they're done). However, just as often the construction actually has a building permit, and it's pretty frequent that the work doesn't even require a permit. I know for a fact that there whole sections of Manhattan where contractors live in fear of one or another curmudgeon with too much time on his or her hands.
In short, 311 does not have the fraudulent-use protections that 911 has, and still gives douchebags a large amount of power. I'm not convinced that adding pics and video to the system will turn the city (any more) into a den of snitches, but the threat is definitely there.
This reminds me of the David Brin stories about a world where senior citizens go around with digital video cameras recording (and streaming to the net) all the time, filming any random group of thuggish-looking young people they see.
@ Thomas "What are the privacy safeguards on the images supplied?"
What prevents someone, right now, from anonymously posting pictures or footage of an emergency or crime to a public forum without obscuring the poeple in the images? Anything? The threat of legal action?
While this seems like a good idea, I have to wonder if this would encourage citizens to violate the privacy of others.
With the current 911 "voice only" system, the person calling 911 is simply reporting a possible crime or other perceived dangerous situation to the authorities who then arrive on the seen to make the "official" assessment.
This would seem to encourage people, who perceive a crime or other dangerous situation, to take "legal" pictures of what today might be considered "illegal" pictures.
As an example (I am sure there are others), a person is walking past a house and hears a noise, looks into an open window in the house and perceives what they might think is a crime being committed in the house. Today, they would call 911 and verbally report what they thought they heard/saw in the house. Under the new system, this person would seemingly be "authorized" to take photos through the open window of the house of the supposed "crime". Now, since this person is using a service that automatically sends all their cameraphone pictures to their Flickr account, these "crime" photos of the inside of the house are now all over the Internet. Previously, such photos would likely be considered "illegal".
Oh, yes, the diagnosis and emergency treatment will go so much better for having received a bunch of grainy, low-res photos.
"Um … help, we have an emergency here …"
"Can you tell me what's happening, Sir?"
"Um … he's bleeding pretty bad … here, let me take a picture … "
"Sir, can you tell me where the injured person is bleeding from?"
"Hold on, what's the email address to send this to?"
"Sir, can you describe the injury to me?"
"Picture's on its way, send an ambulance soon as you can, thanks!"
I hope it doesn't get too many photoshop prank submissions...
As a bicycle commuter, I'd like to be able to send in photos of cars that are illegally parked in the bike lane. As a pedestrian I'd like to be able to send in photos of vehicles extending out of the driveway, completely blocking the sidewalk, making it dangerous for baby strollers, wheelchairs, children on bicycles, etc.
There are a lot of non-emergency things I'd like to able to report with photo, but it's also important to somehow ensure that I haven't modified either the content or date/time. I'm not sure that's possible.
"The worry is that New York will become a city of amateur sleuths and snitches, turning each other in to settle personal scores or because of cultural misunderstandings. But the 911 service has long avoided such hazards. Falsely reporting a crime is itself a serious crime, which discourages people from using 911 for anything other than a true emergency."
In a way it does promote however, the reporting of minor *crimes* that otherwise would be ignored due to the lack of evidence. Therefore the personal revenge against neighbours will start popping up.
"What prevents someone, right now, from anonymously posting pictures or footage of an emergency or crime to a public forum without obscuring the poeple in the images? Anything? The threat of legal action?"
If you're in a public area, absolutely nothing. There would be no grounds for legal action.
Certain commercial activities may have additional restrictions -- by statute, by civil case law, or by self-censorship. For example, not revealing the identity of rape victims or blurring the license plate numbers at car accidents.
Even with facilities like, oh, Bridges and Area-51 you may get harrassed but photography from "public" areas is still legal. If you're taking a lot of photographs, you may even be legitimately investigated whether there is a nefarious reason behind your actions.
Once you enter areas closed to the public, the rules can change -- you can't take photos while you're on Area-51's land. You can be prohibited from photography if you've been invited into a non-public area of a subway system.
I'm not sure were the line is drawn on the open window.
I can't see it being an invasion of privacy if someone has left their drapes open and their house is so close to a public area (i.e. you're not trespassing) that a camera phone takes a decent picture. They didn't have a reasonable expectation of privacy since they didn't take actions to shield their activities from the plain view of the naked eye.
Now, if you need to use telephoto lenses, or you need to setup a step ladder, or take other unusual or extraordinary steps to make it visible from a public area...you've entered a gray area.
As a firefighter, I would love to be able to get a pic of a car accident sent to dispatch (and provided by them to us).
In my rural area, often people from outside the area have no idea where they actually are. Accidents can be called in for a location miles down the road from where it is. We can normally take a look at a photo, and know the location.
Plus it would give us (and the medics with their often-extended response times), a good view of what to expect when we get on-scene.
We are clearly in a grey field here.
Nobody can deny that taking pictures documenting a violation of rights is increasing protection of these rights.
But a lot of this filming and other collecting of biometrics and other data are in themselves violations creating security threats and violating rights of people. Especially if these are networked and/or combined with some sort of automatic recognition or identification the security problems scale dramatically.
This is exactly why the combination between collection and digitalization is so problematic, it scales the security problems exponentially.
It also puts our basic understanding of rights and security requirements of society to a serious test as we need to understand the balance at the source of the problem.
On one side the understandable claim that what is publicly available cannot be considered private and thus anyone can take pictures in the public sphere.
But at the same time we thereby violate the right of a personal sphere in our daily life as we clearly cannot accept a situation in which all people can always assume to be subjected to being filmed in a way that can create threats against their person or assets.
The question is not if we should have a personal sphere somewhere at home where we can withdraw to and have "privacy", but whether this personal sphere does not follow our body EVEN in the public space and thus make ANY attempt for automatic identification and especially collecting biometrics a violation of basic rights and security needs.
It really put to the test what we mean by an open and free society and what it requires for free citizens to uphold control of government and consumer preferences to be enforced in market processes.
Technically we clearly have no need for all these clearly invasive measures such as cameras on open streets since we can solve them with non-invasive and more balanced means combined with an ability of calling defense according to a dynamic threat assesment.
There is a world of difference between our daily life and a situation where we know there is a terrorist with intent and means loose in the underground. But 99% of the people in the underground would participate if they were informed and had the means.
What we do not need is one-sided cures that kill the patient as we mostly see in the extreme anti-terror measures that turns into bigger threats to society than the terror threat itself.
We thus need systems that DON'T provide destabilizing biometrics surveillance and central control in daily operations but can scale according to specific security threats and involve everyone present.
Since the digital world is 100% man-made, of course we can make it. Never underestimate human ingenuity.
The real problem is if we have the social structures to make it happen.
"""What prevents someone, right now, from anonymously posting pictures or footage of an emergency or crime to a public forum without obscuring the poeple in the images? Anything? The threat of legal action?"""
Nothing, quite right.
I'm more worried about a large collection of such material.
I assume it would be saved and archived, and I fear the collection would be protected in line with current best practices (meaning "not very well").
"I'm more worried about a large collection of such material."
Seems to me if a 911 image becomes a part of a medical file (e.g. dispatch sends to a health response team or hospital) then even the smallest collection has to have protection for at least privacy considerations. It's one thing to be the person witnessing an accident and sending in a photo to help with a rescue, and quite another to be someone trying to access stored information from another person's medical files.
I see a serious flaw with the system.
While in phone cameras take reasonable pictures when new after a couple of months sharing a pocket with keys, pens, subway tokens, mints and toffees the quality is lousy.
Even on the very few phones that actually have a lens cover this is still the case it just takes a little longer for the qulity to degrade.
"While in phone cameras take reasonable pictures when new after a couple of months sharing a pocket with keys, pens, subway tokens, mints and toffees the quality is lousy."
I agree that at one end of the spectrum you would have to question the value to the 911 operators of receiving very low quality, low resolution pictures from camera phones.
However, at the other end of the spectrum, there would be those with 400mm zoom lenses on DSLR cameras creating clear, high resolution photos to send to 911 operators.
Since there is no definition/limit of the type of technology used to capture and send the photos to 911 operators, there could be all sorts of technology involved in capturing these "perceived crime" photos, in public and private locations.
The same goes for video that would be sent to 911, while many camera phones can take generally low resolution, low frame rate video, one could also us a HD camcorder with a 100x lense to capture those pictures of a perceived "crime".
Also I would wonder who is the rightful owner of these pictures. Today, the person taking the photo typically is the owner and can do what they want with thier photos (i.e post it on their Flickr account).
If this new direction encourages people to take photos/video of perceived "crimes", in public locations (as well as in private locations?), will there be controls of any sort. I am most worried about photos/video of perceived "crimes" in what are private locations, where today, there are "peeping tom" laws which would make those same photos/videos illegal.
At the "scene of the crime", the responding officer determines it wasn't really an assault, but just an over-zealous couple "having a good time" in the privacy of their home (unfortunately having left a window/shade open). Before, the embarrassed couple would have just apologized, the officer would have a story to tell, and the witness who called 911 would just go home (with their own story to tell). However, now the witness goes home with a bunch of photos/videos of this "crime" that is not really a crime after all.
On a different note, I have to wonder if at some point, the 911 operators would use submitted photos/video to determine the type or urgency of the response which is dispatched to the scene of an accident. I would certainly hope not, especially without anyway to verify the accuracy of the photos/video, but I don't think it can really be ruled out. For example, after making the 911 call and sending photos/videos of an accident scene, the 911 operator, after reviewing the "scene", is allowed to determine the level/type of response. 911 operator: "... I have the photos/video of the automobile accident, it looks like everyone involved is conscious, moving about, and not seriously hurt. I will dispatch this as a Level 2 (medium priority) response ..."
In the 911 response world, those sorts of urgency decisions are already being made.
An auto accident reported with "no injuries" gets shuffled lower than an accident reported with injuries. Just getting a glimpse of the extent of vehicles damage will help properly prioritize accidents.
When the number of indicents reported exceeds the ability of the pool of available units to respond (which does happen on occaision, outside of natural disasters), these sorts of judgements are made.
When 3 people call in at the same time, with a medical problem, they are ranked based on the potential severity of the described symptoms. Someone who's not breathing gets priority of someone having difficulty breathing, whoc gets priority over someone who just lost a finger (which while traumatic, isn't quite as immediately life threatening).
Getting images from those on-scene would greatly help to prioritize the accidents (or just provide vastly more information to the responding units).
on the privacy note, anything sent to 911 is effectively public. Anything you say over hte phone to 911 is recorded, and kept on file. Anything said over the radio between dispatch and fire/ems/police is recorded. I would expect each of these images to be recorded, and filed away as it is now. The amount of paperwork kept by 911 services is impressive. And it's all CYA due to worries about lawsuits.
Posted by: Matt from CT at January 20, 2007 12:01 PM:
>... There are a lot of non-emergency things I'd like to able to report with photo, but it's also important to somehow ensure that I haven't modified either the content or date/time. I'm not sure that's possible.
Well, digital cameras do have "fingerprints": http://www.schneier.com/blog/archives/2006/04/...
How useful and how easily fakable they are is as yet unknown. However, in theory this property could be used as part of a verification mechanism.
It's a far better idea than hundreds of CCTV cams, always on, filming everyone all the time with absolutely no relevance to what you may be doing or saying. After all, if someone decides to send in spurious photos of persons going about their lawful business, they can be traced quite simply and charged with wasting police time.
Seriously, what is the privacy failure mode this idea has that others don't? It doesn't have the problem of CCTV that sweeps up huge amounts of private information without reason, neither does it systematically create vast quantities of irrelevant data to get in the way. It doesn't give the State new powers to apply surveillance to individuals. The only one I can see is that people might submit irrelevant or fake images.
But that should be self-solving. If they are bound to an e164 mobile phone number, they're bound to a billing record, a cellsite-id, and a time-stamp. To abuse someone's privacy you'd have to be at the scene of a crime when it actually happened, be the sole witness, and be able to either a) convince the cops a nonexistent crime occurred or b) convincingly fake a photo in advance to interfere with a real crime.
Good movie-plot threat.
For example, after making the 911 call and sending photos/videos of an accident scene, the 911 operator, after reviewing the "scene", is allowed to determine the level/type of response. 911 operator: ..
Right now they make judgements based purely on verbal description which could much more easily be totally off base. We already deal with that problem. The photo image will not make it worse.
Cool website please vist www.users.nac.net/falken/annoying/
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.