Schneier on Security
A blog covering security and security technology.
« Small Business Identity Theft and Fraud |
| Stealing 130 Million Credit Card Numbers »
August 26, 2009
Interesting video demonstrating how a policeman can manipulate the results of a Breathalyzer.
Posted on August 26, 2009 at 12:20 PM
• 58 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
IIRC, the advice here is to always refuse the breathalyzer and insist on a blood test back at the station. It's less prone to manipulation like this and gives you that much longer to metabolize the alcohol and so have a lower BAC reading.
I believe the math is pretty easy for reversing the effects of metabolism to get the number at the time you were driving, so if you were drunk, they're going to know it, even if you are legal at the time of the sample.
Don't some states make refusing a breath test grounds for a DUI conviction?
@silence: Don't some states make refusing a breath test grounds for a DUI conviction?
I'm not sure what other states do, but here, it is not grounds for a DUI conviction. It's a seperate charge itself and the penalty is loss of license for 3 months.
Maybe it's just me, but I just find it easier to NOT drink alcohol when I am going to drive later.
In NJ you lose your license for refusing a breathalyzer at the scene
Summary for the video-impaired, anyone, please? ...
In the Netherlands, if you fail the breathalyzer test, you get taken to the station for a more accurate follow up test. You're also not allowed to refuse. But even if you were; it seems like a better idea to take the test, because at least then you can be on your way if you pass.
Besides, if you haven't drunk any alcohol at all, then even if the value is multiplied by a hundred, it'll still be zero.
"Summary for the video-impaired, anyone, please? ..."
Test without manipulation = 0.4
Test with manipulation = 0.91
Manipulation consists of putting the finger over the exit hole of the device.
that can backfire - if you've had a drink recently, your BAC will actually increase, not decrease over the short term as the alcohol makes its way into your bloodstream from your digestive system.
A lawyer (who specializes in defending people accused of driving drunk)demonstrates how to manipulate the result of a breathalyzer:
First he explains how police typically hold the breathalyzers, and how it is easy to cover up the hole through which the air leaves the breathalyzer, accidentally or deliberately, when doing so.
He bubbles air through a standardized alcohol solution into a breathalyzer, and produces the correct measurement of .040 (half the legal limit, I think). Then he repeats the process, but this time covers up the exhaust hole of the breathalyzer. Under those conditions, the reading is .091.
I wish the video hadn't been edited in between the two tests. Who's to say the presenter didn't change out the the alcohol solution used for the second test during the lapsed time not present in the video?
I'm not accusing the presenter, just trying to look at it from a lawyer's perspective I guess.
In Romania, AFAIK the law considers the blood test important, the breath is mostly done so that the policeman can have an idea if you're drunk or not. Also, when people refuse to take the test, they're taken to a hospital where a blood test can be done.
Breathalyzers have always been the lowest of the low in junk hocus-pocus technology. Millions of people have been illegally and wrongfully convicted of crimes they didn't commit because of self-righteous pricks jamming these machines up our society's collective ass.
Every time we get a look at the source code, or operating quirks, or pretty much any information about one of these heaps of unconstitutional trash, it's the same: the technology does basically nothing to detect drunk driving.
We'd honestly be getting better justice if the cops simply used a coin flip.
As an interesting note, some diabetics can enter hyperglycemia severely enough that their liver will run the alcohol metabolism cycle in reverse, producing alcohol from excessive blood sugar. In this case, it's possible to have had no drinks, but still blow positive on a breathalyzer.
Of course, you *still* probably shouldn't be driving, but...
I'm skeptical. If the exhaust port is covered over, how is it possible to blow into the device in the first place? Perhaps the idea is that the device is so leaky that with the exhaust port covered over, alcohol-laden breath gets into parts of the device where it shouldn't get? If so, then yes, that's a design defect, but it's not likely to be a defect common to breathalyzers in general.
"We'd honestly be getting better justice if the cops simply used a coin flip."
Probably not. In the UK at least those readers are not used for evidence, calibrated machines back at the station are. The biggest failing is not the technology, it's that alcohol affects people differently. It's hard to make a law like that so we make do with the best of the bad solutions. And as someone above points out, if you don't drink before driving then you don't need to worry about it. (Not that I disagree with disclosing the source code for the machines, I just think there are bigger battles to fight.)
It's not a Breathalyzer, that's a different instrument. It's important to know the difference.
Although it may vary with jurisdiction, readings from portable breath testers are generally not used to convict anyone. They are used to develop probably cause to arrest. Evidence of blood alcohol content comes from another, more sophisticated, instrument.
A lawyer that tells you "don't take a breath test" is making a grab for your wallet. He just wants to rack up billable hours while you try to keep your driver's license.
Finally, the lawyer asserts the device has a "design flaw" that will lead officers to inadvertantly cover the exit port. Look at http://www.ebaumsworld.com/video/watch/80610812/ for a real-life example.
I live in WA state and here if you refuse the test, you will lose your license. I am not sure on what the case law is if you request a blood test instead.
I sat on a DUI case as a jurist a couple of years ago and note some additional facts on the DUI and breathalyzer. They include:
- The law reads, in regards to the breathlyzer, that the test must be within two hours from when you where pulled over.
- The breathlyzer is just one factor toward conviction. Another factor is the way the stat law reads is one is guilty is their driving is impaired (no matter the level--e.g., may be impaired due to other drugs, not just alchohol). In this case, it has to be proven that the driving was impaired. So, the field sobrietry test can come into play here. Also, in the case I was a jurist, the suspect admitted to the cop that he felt his driving was impaired (lesson -- keep ones's mouth shut).
- Interestingly, the suspect wasn't read their rights prior to field sobriety tests.
- Regarding the breathlyzer, after demonstrating the machine, the officer cleaned out the machine, but not the mouthpiece (of the alcoholic mouthwash used to test the machine).
- The most troubling science of the machine, however and not attacked by the defense, was the sampling surrounding the control sample for the machine. No where in the process are the individual control samples tested to ensure accuracy. The batch solution that the sample is drawn from is sampled, but not the individual sample units. The scientists involved need to read what Dr. Deming wrote regarding batch sampling. I am surprised it hasn't been attacked.
A lawyer to whom I'm related, who practices in a distant state (so while I don't trust him completely, he at least doesn't have an interest in my incarceration), tells me not even to roll down my window if I'm drunk when a cop pulls me over. Just open the window a quarter of an inch and slide my license and insurance info out the slot.
I'll probably get beaten and my car impounded, and I'll have to hire a lawyer, but it will be quite difficult for them to convict me or take my license in this scenario. If I submit to a breath test, whether I'm drunk or not, my fate is out of my hands, no matter who I hire.
The reason I don't do this when I haven't been drinking, although it would be within my rights, is to convince the cop not to breathalyze me in the first place. Then I can avoid the beatings and impoundings and attorney fees too.
Out of interest, why did he say "if you're over 21 don't give a breath test"? What's the difference under 21?
What remains unsaid is that the manipulation by the officer can provide the PC for arrest. If one is at 0.04 and the officer causes him to blow over the limit, then that's a trip to the station. It's a tool of legal battery and statistic generation: you never see any news reports about how many of the Labor Day DUIs resulted in convictions.
The officer is supposed to wait 15 minutes after the subject eats, vomits, or puts anything in the mouth.
We've all seen cops standing on the roadside, waiting patiently.
Nope, never happens.
And the manufacturers won't reveal trade secrets, such as the specificity of the device.
Good luck trying to cross-examine that witness.
Nonny Bunny, did you leave out a zero? At .4 you may well be dead, at .91 it's really freaky if you aren't.
The reason for the edit between tests is likely that the device takes time to clear after before a second test can be run. That is over 10 minutes and the lawyer probably didn't want to vamp for that period on a video.
I guess the point of refusing is that you avoid incrimination based on this type of device and can then appeal the other charges filed.
Wonder if anyone has thought of just carrying their own breath device and using it as a defense, similar to this guy who setup his own cam to catch abuse by police
"the device takes time to clear...over 10 minutes"
Where do you get 10 minutes? The device specs from the manufacturer say 15 seconds between tests:
"Intervals before readiness for operation:
1st measurement after determining a previous test result in range: approx. 15 s
up to 0.19 mg/L: 10 s
> 0.95 mg/L: 90 s"
I also noticed an odd calibration requirement.
"Calibration interval recommended 6 months (depends on required accuracy)"
What is considered "required" accuracy and how that is audited.
Just as every cop is a criminal and all the sinners saints ...
"Nonny Bunny, did you leave out a zero? At .4 you may well be dead, at .91 it's really freaky if you aren't."
I may have. The device shows 040 and 091. I don't know what units, and it doesn't clearly show a decimal point (but it might be to the far left)
"Besides, if you haven't drunk any alcohol at all, then even if the value is multiplied by a hundred, it'll still be zero."
Not so. Breathalyzers are notoriously prone to false positives (e.g. due to eating some brands of bon-bons). This is a good way to amplify them.
@Ben in AZ "a lawyer's perspective "
It's true you can't always trust the video's you see on the internets. See From Mythbusters ep 118 "YouTube Special".
Let's choose a different perspective then Ben.
How about science? Seems a simple enough test to replicate.
If it can be repeated then the technology is faulty, prone to error, and people shouldn't be convicted based on it's results. If the deputy or trooper manipulated the device then they've lost their PC and any evidence afterwards should be disallowed.
I am not sure I count the video as solid proof for the reasons others have stated here, but it would explain the impossibly high numbers you occasionally see on the internet. Recently, I saw a report about a woman with a .044. I find it hard to believe she was driving, walking or doing much of anything. If she really had a .022 it would be in line with what we are told to expect with that level of intoxication.
I think the charge should be 'driving poorly' not DUI/DWI and then it would not matter what the BAC is. If someone cannot keep in a lane, I don't care if they are tired, drunk or stoned they should get a ticket.
Isn't the bigger story here that someone actually used bing?
I can confirm that in my state the handheld devices are used in order to determine whether to take someone in for the more accurate test which can be used in court.
My brother is a law enforcement officer and he's told me stories of people who were drunk while driving or on work release or home detention, etc, but were arrested in such a rural area that they were under the limit by the time the cruiser got back to town.
Apparently the courts don't do the math to calculate what the BAC must have been 45 minutes ago in this area.
It would probably just confuse the jurors and result in a mistrial anyway.
DUI laws, test protocols, test devices, administrative suspension rules etc. all vary from state to state. Although this lawyer's advice might be good in his state, it might be really bad advice in yours.
Before the Europeans get high-and-mighty: Most US cities have no adequate public transportation, and are zoned in such a way that there are no "local" bars in most residential areas, so walking is out of the question. Therefore, the only way to get to or from a bar is by driving.
Result? Nearly everyone drives after drinking, and most are somewhere near the legal limit, though still short of being drunk (the limits are very low).
To compound the problem, our police forces are parasitic--they fund themselves by issuing traffic tickets and fines. 99% of violent crimes and thefts are just recorded and filed away. They don't make money for the cops, so they aren't investigated.
The result is a system where the police hunt peaceful, responsible citizens who are driving home from bars after two drinks, while people are being raped nearby but get no help from the cops. Defending against DUI convictions is a big business here.
nick I agree with your descriptions of American law enforcement but I imagine residents of most civilized European nations might have a chuckle at our drunk drivers, at our public transport, AND at our brutally inept police.
nick, if you were a cop, would you rather be busy hassling peaceful, responsible citizens or armed pcp-charged criminals? The invisible hand behind our reward system says ignore the dangerous ones.
nick, Jess, Dave X
I agree with you that bar locality situation is much better here in Europe but police is not much better if at all. Often it is worse then in the USA, especially in the eastern parts of the continent. As probably everywhere, thay go after easy prey.
BTW. I had only one encounter with US highway patrol during the total 4 weeks there and the officer has been very nice, only requesting me to go 75 mph not 77 (drat that cruise control). Maybe I was just lucky :-)
"Breathalyzers are notoriously prone to false positives (e.g. due to eating some brands of bon-bons)."
You mean the type of bon-bons with alcohol in them? I wouldn't call that a false positive. It would be a false positive if it measured something other than alcohol, not if it measured alcohol from a different source than drinks.
@Hammy G "Isn't the bigger story here that someone actually used bing?"
no that that person was Bruce S.
The video seems to be falsified. The time code jumps from 9/28/07 04:53 to 9/28/07 05:09. See 1:22-1:25 in the video. I smell hokey.
Jess, would you really rather be beaten seriously by one or more strangers than risk losing your driver's license?
Your kink is okay, but don't expect them to respect your safeword.
In Australia, if you fail or refuse a breathalyser test your reward is a mandatory blood test, which is done at a hospital - and you get one of the two samples that are taken at the same time.
If that comes out negative, nothing else happens.
It's the blood test that is used in court - the breathalyser is an indicator that more accurate evidence is required. If police got a lot of false positives from their breathalysers, they'd be looking at why that was and either making the threshold on the breathalyser higher, or retraining to avoid this sort of mistake.
There's no point to them manipulating the breathalyser results, as it's wasting their time, and our money.
Actually I just thought of a point for that manipulation - if they thought that you were driving under the influence of marijuana, which they don't have a roadside test for, then this might be an easier way to get a blood test done.
Not quite true.
If you refuse a preliminary breathtest in Australia, that is a separate offence with penalties commensurate with recording a high-range blood alcohol concentration.
If you fail the PBT, (ie read >0.04, or >0.01 if on a restricted licence), then you are required to submit for further analysis which is usually a formal breath-test on a large machine at the police station (or in the 'booze bus'). If you have a reason that cannot be done (eg asthma), a blood test is the only other option, which is performed by a doctor and you get one of the THREE samples taken at the same time. One is for 'screening', and if it is positive, it is confirmed by analysis of the 'police' sample. The third sample goes with the offender and can be independently analysed by them for evidentiary purposes.
So, in Australia, the reading of a hand-held breathalyser is *never* enough to convict.
"Maybe it's just me, but I just find it easier to NOT drink alcohol when I am going to drive later."
This is an emotional response. The question posed is not whether people should be trying to fool the system, but whether the system CAN be fooled.
An analog is found in the debate as to whether new security vulnerabilities should be made public when a corporation who produces a security product refuses to acknowledge (or fix) the vulnerability in a reasonable time frame. Often enough, someone will post a comment along the lines of "it's unethical to release this publicly because you're only making it easy to break security, are you on the side of law breakers?" The motivation of the author is usually to spur a rapid solution to the problem, which will protect people who routinely patch security (which is advised by many security experts).
This is a common, recurring theme in security, and your variation on the typical response is equally unhelpful. Hopefully I've done a good job illustrating that.
"The video seems to be falsified. The time code jumps from 9/28/07 04:53 to 9/28/07 05:09. See 1:22-1:25 in the video. I smell hokey."
Huh? If the video was edited very well and there was an attempt to conceal that fact, I'd be inclined to agree.
Since it's obvious that the video has been edited, I'll go with Occam's razor until I see compelling evidence that the editing was done for nefarious purposes. In this case, the simplest answer would be that the editing was done to make a point quickly for a broad audience (many of whom probably have attention span issues).
Besides, he says that he specializes in the defense of people accused of DUI. It follows that he probably has to demonstrate this effect, in real time, to a jury.
From a scientific standpoint, it makes perfectly good sense that covering the port on the back of the tube would concentrate the amount of alcohol in the device. The apparatus he is using to demonstrate this effect almost certainly would have to be using the venturi effect: blowing perpendicular to the tube which contains the alcohol solution will take some of the solution with it. The same physics principle is demonstrated often enough by nature in mountainous regions where a crosswind and a gap between two mountains creates enough force to jackknife, or even topple, semi trucks.
Covering the port probably also imparts a bit of body heat from the examiner's finger, and slowing the flow of breath through the tube will impart more warmth from the examinee's breath to the inner walls of the tube and the device itself. Warmer air holds more water (and more alcohol), and causes more alcohol to evaporate from the solution in the testing apparatus, concentrating more of it in the air.
A lot of demostrable physics principles are in effect here, so I have no particular reason to doubt the authenticity of the test I observed him perform.
I Googled "warmer air" "more alcohol" to see if others made similar arguments.
One of the results:
This discusses the effects of temperature on alcohol in breathalyzer tests. It's focus is different from the video Bruce linked (techniques an examinee may employ to try and beat the test, as opposed to techniques that a law enforcement officer may employ to obtain a higher likelihood of eventual conviction).
tc, you stated that the defendant admitted to the cop that he felt his driving was impaired. Did you have evidence other than the cop's word for that?
I was a passenger in my boyfriends car and we went through a check point. He passed the sobriety test yet they still arrested him and took him to get his blood drawn. They took me out of the car and breathalyzed me numerous times. After the 4th or 5th time they finally stopped and took me to where he was. The office approached me again and asked me to breathalyze me again and i refuse. What should i be expecting to hear when they apparently charged me with an underage. They never told me my B.A.L and are they allowed to tell me to take the test numerous times? What if i said no?
Yeah, I don't see the fuss here. If you've been drinking alcohol, you ABSOLUTELY should not be driving. Playing games with being a few points below the technical limit and contesting the breathalyzer results is, in a word, sociopathic. The limit is set high deliberately so that there's absolutely no chance somebody can be convicted of drunk driving because they had a tablespoon of cough syrup or a bit of cherries jubilee that didn't quite burn all the alcohol off. If you are anywhere near being within an order of magnitude of the limit, it's because you've been drinking actual alcoholic beverages, and if you've been doing that you have absolutely no business getting behind the wheel of a motor vehicle, period. Call for a ride, or walk.
Driving motor vehicles on the public roads is dangerous enough without having people try to do it in an impaired state and while sending text messages on a cell phone.
Jonadab, you are obviously a teetotaler, I am in my seventies and have been drinking all my adult life, I have never had a DUI or even an accident when drinking, some of us know how to drink and when to stop.
P.S. I drive about 20,000 miles per year.
Ron, you are obviously an idiot and very lucky. The evidence of the effect of alcohol on reaction times is probably the least controversial of all the points raised here, and I would love to see you try and argue this with one of the many families who lose loved ones due to drivers under the influence.
Why did you post a link to Bing?
Why didn't you just link to the video? It is clearly posted on YouTube and the link to it is clearly displayed underneath the video on the Bing page. Why link to Bing?
One year later, the Bing search result link is completely useless to locate the video. Numerous other searches of mine have failed to find it. Can anyone relocate this video, preferably with a permalink?
Not sure why you need the actual video (it appears to be as described above from my memory of it)
The issue with most "toy testers" used for motoring offences (speed, blood alcohol drugs etc) is they are generally easy to manipulate by the person taking the test and they don't have duplicate evidence which can be tested independantly (it's why you used to get given a second sample of blood).
For instance how often have you seen or even heard of the Police waiting the 15mins observation time required to ensure there is no contamination from things like vomit or belches or hic upps?
Compounding this are the laws surrounding the use of these toy testers... they are so badly written that you can "be guilty" simply by asking what your rights are (ie an officer can "misconstrue" that as a refusal and thus you are guilty of another offence etc).
The best advice is know your state laws in advance say next to nothing and know the number of a lawyer specialising in the states DUI laws...
As for manipulating the device by putting your finger over the exit port yes it does effect the reading as does having a very cold mouth piece.
Basicaly there is an assumption in the software of what the temprature of the equipment and ambient temprature are, alcohol has a non linear condensing rate which is quite wide at typical ambient tempratures. Thus by making you blow hard and slow both the apparent and real alcohol rate rises.
Have a look at,
And also the NCDD site it links to,
I know one or two people that have practiced law in the US and UK who have the spoken attitude that once a police officer pulls you over you can be found guilty of something you can be convicted of. This is irrespective of what you may or may not have done and is simply due to the legislation available to the police including laws well over three hundred years old in the UK...
300 years old? Did they have cars in England 300 years ago? You'd think they could have designed ones that don't leak oil, like the Japs did.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.