[Brinworld] Car's data recorder convicts driver
(ok, from slashdot..) http://www.newhouse.com/archive/jensen061203.html
On Monday June 16 2003 09:59, Major Variola (ret.) wrote:
(ok, from slashdot..) http://www.newhouse.com/archive/jensen061203.html
I personally find the privacy implications of EDRs rather unsettling. This story doesn't change that one bit. However, in this particular case, I don't think what the EDR said really matters. The three paragraphs from the story say a lot about what happened here: | Matos was driving the 2002 Pontiac Trans Am in a 30 mph zone of a | suburb near Fort Lauderdale, Fla., when the car driven by a teenage | girl pulled out of a driveway into his path. | | The driver and her friend died instantly. | | Defense lawyer Robert Stanziale said Matos was going about 60 mph. | Assistant State Prosecutor Michael Horowitz said that his accident | investigator calculated Matos was traveling about 98 mph. The | electronic data recorder in Matos' car showed his peak speed was 114 | mph in the seconds before the crash. The *defense* attorney said his client was going 30 mph over the limit (60 mph in a 30 mph zone)! That is a grossly inappropriate speed in a residential area. Here in Texas, a ticket for 55 mph in a 30 mph zone cannot be dismissed with DSC. Not sure how the law works in Florida but I would be surprised if it was that dissimilar. Let's assume for the moment the prosecution's accident invesitigator is totally full of bovine excrement, and that all manner of gremlins snuck into the EDR thus causing it to record a grossly inaccurate peak speed, and thus, the only version of the story we can give full credibility to is the defense's version. If I were on that jury, I'd still vote for a conviction. Matos is a scofflaw and deserves exactly what he is getting. -- Shawn K. Quinn
At 11:16 AM 6/16/03 -0500, Shawn K. Quinn wrote: ...
I personally find the privacy implications of EDRs rather unsettling. This story doesn't change that one bit. However, in this particular case, I don't think what the EDR said really matters. The three paragraphs from the story say a lot about what happened here:
... It seems intuitively like the EDR ought to be about as valuable to the defense as the prosecution, right? E.g., the prosecutor says "this guy was driving 120 miles an hour down the road while being pursued by the police," but the EDR says he'd never topped 70. There are creepy privacy implications in there somewhere, but the basic technology seems no more inherently Orwellian than, say, DNA testing--which seems to be a pretty good way of actually locking up the right guy now and then, rather than someone who looks kind-of like the guy who did it, and was seen in the area by an eyewitness and picked out of a police lineup. ...
Shawn K. Quinn
--John Kelsey, kelsey.j@ix.netcom.com PGP: FA48 3237 9AD5 30AC EEDD BBC8 2A80 6948 4CAA F259
Seems like a market for "open source" EDRs could be a good one. A user accessible reset button could come in handy. steve
On Tuesday, June 17, 2003, at 03:48 PM, Thomas Shaddack wrote:
On Mon, 16 Jun 2003, Steve Schear wrote:
Seems like a market for "open source" EDRs could be a good one. A user accessible reset button could come in handy.
Could a stun gun help?
Unlikely. Getting juice into the innards of a box in a way so as to overwrite data is not nearly so simply as applying sparky things to the outside of the box. Lots of reasons for this. --Tim May
On Tue, 17 Jun 2003, Tim May wrote:
Unlikely. Getting juice into the innards of a box in a way so as to overwrite data is not nearly so simply as applying sparky things to the outside of the box. Lots of reasons for this.
The idea wasn't about overwriting the data. The idea was about frying the chip with the data inside (and if all the other chips inside the box become a collateral damage, let's that be so). As long as it is outside the technological abilities of the given adversary to retrieve the data from the fried chip, the objective is reached. The idea also wasn't about the outside of the box, I thought rather disconnecting the power leads and blasting the spark into the power-GND pair, or into the (disconnected, we don't want to kill the entire car electronics) data bus. With a bit of luck, the spark could get through the filters and into the Vcc pins of the chips.
Why go to all that trouble. Just take it out of circuit. Cut the printed circuit board leads and disable it or if its in an inaccessible black box, cut the leads to the box. Easy enough. On Wednesday, June 18, 2003, at 04:11 PM, Thomas Shaddack wrote:
On Tue, 17 Jun 2003, Tim May wrote:
Unlikely. Getting juice into the innards of a box in a way so as to overwrite data is not nearly so simply as applying sparky things to the outside of the box. Lots of reasons for this.
The idea wasn't about overwriting the data. The idea was about frying the chip with the data inside (and if all the other chips inside the box become a collateral damage, let's that be so). As long as it is outside the technological abilities of the given adversary to retrieve the data from the fried chip, the objective is reached.
The idea also wasn't about the outside of the box, I thought rather disconnecting the power leads and blasting the spark into the power-GND pair, or into the (disconnected, we don't want to kill the entire car electronics) data bus. With a bit of luck, the spark could get through the filters and into the Vcc pins of the chips.
On Wed, 18 Jun 2003, jburnes wrote:
Why go to all that trouble. Just take it out of circuit. Cut the printed circuit board leads and disable it or if its in an inaccessible black box, cut the leads to the box.
Easy enough.
Works very nicely. :) Problem: leaves evidence, and takes time. The main advantage of electric shock is that the fried chip looks for the naked eye exactly the same way as a non-fried chip. The only difference could be found with a scanning electron microscope on the chip itself, which is something nobody is likely to bother with. Especially in harsh environments (cars classify) chips tend to die, so its death could look as natural enough to not be suspicious. If I am wrong, please tell me where and why. :)
On Thu, 19 Jun 2003, Thomas Shaddack wrote:
Works very nicely. :)
Problem: leaves evidence, and takes time. The main advantage of electric shock is that the fried chip looks for the naked eye exactly the same way as a non-fried chip. The only difference could be found with a scanning electron microscope on the chip itself, which is something nobody is likely to bother with. Especially in harsh environments (cars classify) chips tend to die, so its death could look as natural enough to not be suspicious.
If I am wrong, please tell me where and why. :)
Automotive environments are known to be harsh, so electronics is protected to some extent. The assumption is that spark plug voltages can get into sensors, so most data lines are protected as are the sensor lines. If you try to fry things with double the voltage of a standard spark plug it may not work, if you use 10 times that it will, but the ESD protection will obviously be blown too. That begins to look suspicious (but I doubt anyone could _prove_ you fried it on purpose). The main question is how deep is the memory of these things. If they only remember the last catastrophic event then "privacy" isn't a problem. The actual routes taken are not stored. If you are in an accident and the cops ask you to take a breath test, you can take the test or not - and deal with the consequences of the legal system based on your choice. The data taken from the recorder for the "event" is then corroboration, which may help instead of hurt you. If the box remembers everything you do, and the garage mechanic can use it to blackmail you, then it becomes a "privacy" issue. I think the issue is when data is removed, and how much is actually stored. Can anybody explain the details? Patience, persistence, truth, Dr. mike
Mike Rosing <eresrch@eskimo.com> wrote:
Automotive environments are known to be harsh, so electronics is protected to some extent. The assumption is that spark plug voltages can get into sensors, so most data lines are protected as are the sensor lines. If you try to fry things with double the voltage of a standard spark plug it may not work, if you use 10 times that it will, but the ESD protection will obviously be blown too. That begins to look suspicious (but I doubt anyone could _prove_ you fried it on purpose).
In automotive power systems, the bigger concern is load dump. When there's a step change in alternator load from high to low, the commutating inductances resulting from the field windings of the alternator can't react quickly enough, and you tend to get big spikes on the power lines. Perhaps this could be used to advantage---if you want to convince someone that your electronics blew up on their own, blow up the rectifier at the output of your alternator, too. -- Riad Wahby rsw@jfet.org MIT VI-2 M.Eng
On Wednesday, June 18, 2003, at 06:15 PM, Thomas Shaddack wrote:
On Wed, 18 Jun 2003, jburnes wrote:
Why go to all that trouble. Just take it out of circuit. Cut the printed circuit board leads and disable it or if its in an inaccessible black box, cut the leads to the box.
Easy enough.
Works very nicely. :)
Problem: leaves evidence, and takes time. The main advantage of electric shock is that the fried chip looks for the naked eye exactly the same way as a non-fried chip. The only difference could be found with a scanning electron microscope on the chip itself, which is something nobody is likely to bother with. Especially in harsh environments (cars classify) chips tend to die, so its death could look as natural enough to not be suspicious.
If I am wrong, please tell me where and why. :)
The point being that sensor data from outside the box does NOT get written to either flash or disk drive storage directly. It is collected from many places and fed through the assortment of microprocessors. High voltages are clamped in the usual ways, with Schottky diodes protecting the inputs, etc. Even if signals massively outside the specs got into the boxes, it would be the processors which got fried, not the storage devices. This was my point about how "sparky things" would not overwrite data. It takes logic to correctly write to storage. The processors and peripheral logic _might_ be zapped, but the storage chips would almost certainly not have been erroneously overwritten...just a matter of disconnecting them and reading them in another system, something most forensic or recovery labs probably have many jigs set up for. --Tim May
On Mon, Jun 16, 2003 at 05:11:57PM -0400, John Kelsey wrote: ...
It seems intuitively like the EDR ought to be about as valuable to the defense as the prosecution, right? E.g., the prosecutor says "this guy was driving 120 miles an hour down the road while being pursued by the police," but the EDR says he'd never topped 70. There are creepy privacy implications in there somewhere, but the basic technology seems no more inherently Orwellian than, say, DNA testing--which seems to be a pretty good way of actually locking up the right guy now and then, rather than someone who looks kind-of like the guy who did it, and was seen in the area by an eyewitness and picked out of a police lineup.
The types of problems with DNA testing such as state's refusal to allow testing of convicts when it might prove their innocence, and testing lab "errors", would also apply to EDR boxes. I.e. states will contrive to use EDR records only when it proves their case, and data recovered will be subject to "interpretation". You can bet that when EDRs become important as evidence, citizens won't be allowed to posess the means to read their own EDRs let alone write to them. Eric
On Mon, Jun 16, 2003 at 05:11:57PM -0400, John Kelsey wrote: | At 11:16 AM 6/16/03 -0500, Shawn K. Quinn wrote: | ... | >I personally find the privacy implications of EDRs rather unsettling. | >This story doesn't change that one bit. However, in this particular | >case, I don't think what the EDR said really matters. The three | >paragraphs from the story say a lot about what happened here: | | ... | It seems intuitively like the EDR ought to be about as valuable to the | defense as the prosecution, right? E.g., the prosecutor says "this guy was | driving 120 miles an hour down the road while being pursued by the police," | but the EDR says he'd never topped 70. There are creepy privacy | implications in there somewhere, but the basic technology seems no more | inherently Orwellian than, say, DNA testing--which seems to be a pretty | good way of actually locking up the right guy now and then, rather than | someone who looks kind-of like the guy who did it, and was seen in the area | by an eyewitness and picked out of a police lineup. Just wait 'till they integrate GPS, and GPRS or 802.11. Much of this can be seem in the OnStar systems, which haven't yet featured in divorce proceedings, afaik. You can call up and find out where your car is. Adam PS: Bob Blakely once defined privacy as the right to lie and get away with it, which fits into some of what many people mean by privacy. -- "It is seldom that liberty of any kind is lost all at once." -Hume
Just wait 'till they integrate GPS, and GPRS or 802.11.
Transmitter is easy to find. Receiver is easy to jam with a micropower jammer. Sometimes all you need could just be creatively tweaking the ignition and antenna wiring to get "faulty shielding" in the right places; it requires much more experience to make it look "accidental", though.
Much of this can be seem in the OnStar systems, which haven't yet featured in divorce proceedings, afaik.
Matter of time. The next generation of sleuths will be much more tech savvy than the current one.
You can call up and find out where your car is.
...eg, in a nameless radio shadow.
Adam
PS: Bob Blakely once defined privacy as the right to lie and get away with it, which fits into some of what many people mean by privacy.
Another possible definition is the right to tell the truth and get away with it. But both definitions are rather about free speech than about privacy, but then we'd get to a fight over definitions which is in this context better to leave on the shoulders of people making encyclopedias.
On Tuesday, June 17, 2003, at 03:48 PM, Thomas Shaddack wrote:
Adam
PS: Bob Blakely once defined privacy as the right to lie and get away with it, which fits into some of what many people mean by privacy.
Another possible definition is the right to tell the truth and get away with it.
But both definitions are rather about free speech than about privacy, but then we'd get to a fight over definitions which is in this context better to leave on the shoulders of people making encyclopedias.
Maybe I have a minor corollary to Somebody's Law: "All debates about privacy eventually degenerate into foolish and off-target debates about the meaning of truth." It never makes sense to argue about a "right to lie" or a "right to tell the truth." One man's lie is another man's truth. And even _asking_ for a true response is usually an overstepping, as it presumes the asker knows what is true and what is not. Pilate said it all 2000 years ago. --Tim May
On Tue, Jun 17, 2003 at 09:30:35PM -0700, Tim May wrote: | On Tuesday, June 17, 2003, at 03:48 PM, Thomas Shaddack wrote: | > | >>Adam | >> | >>PS: Bob Blakely once defined privacy as the right to lie and get away | >>with it, which fits into some of what many people mean by privacy. | > | >Another possible definition is the right to tell the truth and get away | >with it. | > | >But both definitions are rather about free speech than about privacy, | >but | >then we'd get to a fight over definitions which is in this context | >better | >to leave on the shoulders of people making encyclopedias. | > | | Maybe I have a minor corollary to Somebody's Law: "All debates about | privacy eventually degenerate into foolish and off-target debates about | the meaning of truth." | | It never makes sense to argue about a "right to lie" or a "right to | tell the truth." One man's lie is another man's truth. And even | _asking_ for a true response is usually an overstepping, as it presumes | the asker knows what is true and what is not. Pilate said it all 2000 | years ago. I wasn't arguing, I was quipping. I find the many meanings of the word privacy to be fascinating. So when someone commented that the car's tattle-box is or isn't a privacy invasion, I thought I'd offer up a definition under which it is. Its a definition that lots of people use, as John points out. Perhaps better than 'right' would be 'ability,' 'The ability to lie and get away with it.' Adam -- 'No, honey, I was working late at the office.'
On Wed, 18 Jun 2003, Adam Shostack wrote:
I wasn't arguing, I was quipping.
I find the many meanings of the word privacy to be fascinating. So when someone commented that the car's tattle-box is or isn't a privacy invasion, I thought I'd offer up a definition under which it is. Its a definition that lots of people use, as John points out.
Perhaps better than 'right' would be 'ability,' 'The ability to lie and get away with it.'
Adam
-- 'No, honey, I was working late at the office.'
Reminds me of the first time I saw a guy with a "brick phone". I'm in a bar eating a burger and drinking beer, and this guy sits down one seat away from me, pulls out this huge cell phone, and starts punching away. 10 seconds later he's saying "I'll be late coming home, I have more work to do in the office". Like she can't here the background music! Oh well, I was getting paid for his air time :-) Patience, persistence, truth, Dr. mike
At 08:17 2003-06-18 -0400, Adam Shostack wrote:
| It never makes sense to argue about a "right to lie" or a "right to | tell the truth." One man's lie is another man's truth. And even | _asking_ for a true response is usually an overstepping, as it presumes | the asker knows what is true and what is not. Pilate said it all 2000 | years ago.
I wasn't arguing, I was quipping.
I find the many meanings of the word privacy to be fascinating. So when someone commented that the car's tattle-box is or isn't a privacy invasion, I thought I'd offer up a definition under which it is. Its a definition that lots of people use, as John points out.
Perhaps better than 'right' would be 'ability,' 'The ability to lie and get away with it.'
Indeed 'privacy' and 'secrecy' are often confused and their meanings overlap in many a mind. I think that most, at least in the West, accept that privacy "..is based on rules and trust," for example, records kept on us by our doctors. Because exposure of various aspects of our private lives can do lasting damage, privacy is only effective when controlled by the party seeking it, who may disclose it or not as they see fit and can only be guaranteed when those who would "sell you out" don't possess the possibly damaging information. For that reason among others, I am really only interested in privacy mediated by personal secrecy and technologies I trust and/or control. steve
On Wed, 18 Jun 2003, Steve Schear wrote:
Indeed 'privacy' and 'secrecy' are often confused and their meanings overlap in many a mind. I think that most, at least in the West, accept that privacy "..is based on rules and trust," for example, records kept on us by our doctors. Because exposure of various aspects of our private lives can do lasting damage, privacy is only effective when controlled by the party seeking it, who may disclose it or not as they see fit and can only be guaranteed when those who would "sell you out" don't possess the possibly damaging information. For that reason among others, I am really only interested in privacy mediated by personal secrecy and technologies I trust and/or control.
I agree with you. Being anonymous is very important here. Privacy is something alluded to by the famous "Gentlemen do not read other gentlemen's mail". Secrecy is what other people cannot find out. Anonymity (strong or not) is vastly important to secrecy. Medical data is a great example of this. It may be private, for some (weak) values of private, right now. Being John Doe at the doctor's office and paying cash, though, is vastly better in terms of assurance, at least until the doctor's business-cam interfaces with other databases. Too bad that works so poorly with insurance, but then worker insurance in the US is nearly a government program, anyway. -j -- Jamie Lawrence jal@jal.org A computer without a Microsoft operating system is like a dog without bricks tied to its head.
At 11:45 2003-06-18 -0500, Jamie Lawrence wrote:
Anonymity (strong or not) is vastly important to secrecy.
Medical data is a great example of this. It may be private, for some (weak) values of private, right now. Being John Doe at the doctor's office and paying cash, though, is vastly better in terms of assurance, at least until the doctor's business-cam interfaces with other databases. Too bad that works so poorly with insurance, but then worker insurance in the US is nearly a government program, anyway.
There may be a viable opportunity for an off-shore private medical insurance carrier which does not use your social security number as your identifier to the medical service provider. Due to excessive U.S. fed and state insurance regulations many/most doctors might refuse to accept it (at least initially) it may be necessary for this insurance to operate "off network" so that subscribers would have to pay the care giver and be reimbursed. steve
On Wednesday, June 18, 2003, at 05:17 AM, Adam Shostack wrote:
I wasn't arguing, I was quipping.
I find the many meanings of the word privacy to be fascinating. So when someone commented that the car's tattle-box is or isn't a privacy invasion, I thought I'd offer up a definition under which it is. Its a definition that lots of people use, as John points out.
Perhaps better than 'right' would be 'ability,' 'The ability to lie and get away with it.'
I wasn't picking on you or your points, that's for sure. In fact, I barely noticed whose message I was replying to. My point was a larger one, that nearly all such debates about privacy eventually come round to issues of "what have you got to hide?" and issues of truth and lies. This is why I like the "Congresss shall make no law" and "shall not be infringed" absoluteness of the original Constitution. The language does not natter about "truthful speaking shall not be infringed." And this is why more recent legislation allowing government to regulate "commercial speech" or to decide which speech is true and which is false (as in advertising claims) is so corrosive to liberty. --Tim May "The great object is that every man be armed and everyone who is able may have a gun." --Patrick Henry "The best we can hope for concerning the people at large is that they be properly armed." --Alexander Hamilton
On Wed, Jun 18, 2003 at 09:11:58AM -0700, Tim May wrote: | On Wednesday, June 18, 2003, at 05:17 AM, Adam Shostack wrote: | > | >I wasn't arguing, I was quipping. | > | >I find the many meanings of the word privacy to be fascinating. So | >when someone commented that the car's tattle-box is or isn't a privacy | >invasion, I thought I'd offer up a definition under which it is. | >Its a definition that lots of people use, as John points out. | > | >Perhaps better than 'right' would be 'ability,' 'The ability to lie | >and get away with it.' | | I wasn't picking on you or your points, that's for sure. In fact, I | barely noticed whose message I was replying to. Gives new meaning to anonymous postings. ;) | My point was a larger one, that nearly all such debates about privacy | eventually come round to issues of "what have you got to hide?" and | issues of truth and lies. | | This is why I like the "Congresss shall make no law" and "shall not be | infringed" absoluteness of the original Constitution. The language does | not natter about "truthful speaking shall not be infringed." | | And this is why more recent legislation allowing government to regulate | "commercial speech" or to decide which speech is true and which is | false (as in advertising claims) is so corrosive to liberty. Indeed. The European data protection laws are fundamentally unamerican. Unfortunately, Congress has made laws, numbering each of us, and then tries to regulate the abuse of that (free, freely usable, legally enforced) numbering scheme. Adam -- "It is seldom that liberty of any kind is lost all at once." -Hume
participants (12)
-
Adam Shostack -
Eric Murray -
Jamie Lawrence -
jburnes -
John Kelsey -
Major Variola (ret.) -
Mike Rosing -
Riad S. Wahby -
Shawn K. Quinn -
Steve Schear -
Thomas Shaddack -
Tim May