Schneier on Security
A blog covering security and security technology.
« The FBI and Wiretaps |
| Zero-Tolerance Policies »
November 3, 2009
Detecting Terrorists by Smelling Fear
The technology relies on recognising a pheromone - or scent signal - produced in sweat when a person is scared.
Researchers hope the ''fear detector'' will make it possible to identify individuals at check points who are up to no good.
Terrorists with murder in mind, drug smugglers, or criminals on the run are likely to be very fearful of being discovered.
Seems like yet another technology that will be swamped with false positives.
And is there any justification to the hypothesis that terrorists will be more afraid than anyone else? And do we know why people tend to feel fear? Is it because they're up to no good, or because of more benign reasons -- like they're scared of something? This link from emotion to intent is very tenuous.
Posted on November 3, 2009 at 6:12 AM
• 75 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
what about people that are e.g.
- afraid of flying
- meeting their in-laws
- forgot to lock their home and just remembered it
- have an important business meeting
there are many occasions where the human brain trigger a 'fear' response other than being a 'bad guy'
(of course not talking about myself here)
What if I'm afraid the system will falsely think I'm a terrorist because I'm afraid the system will falsely identify me as a terrorist?
I had a long conversation with a man from the UK Border Agency a little while ago; he said the single most effective weapon in the arsenal against malefactors is experience and instinct.
No, it's not foolproof but 99 times out of a hundred it's better than a machine or a computer.
Why are we continually trying to find a technological way to look into peoples' souls to see there intention when we instinctively know that reading humans and their intentions is something that intelligent well trained humans do better than machines?
It's taken us enough time and money to build a computer that can beat the best human at chess; and chess is very simple in comparison to reading minds.
Not to mention the fact that terrorists probably experience _less_ fear than the average person.
This is shown scientifically at least for psychopaths.
@uk visa :
The justification is that, in the long run, it's faster and cheaper.
And no, I don't think it's a good idea either.
This seems to be based on the theory that "If you've nothing to hide, you've nothing to fear".
What if I'm scared because of the constant fearmongering?
Who aren't afraid?
People doing good work and are righteous in the eyes of the almighty.
There's no evidence that terrorists are a class of psychopaths or sociopaths for that matter.
@uk visa "99 times out of a hundred it's better than a machine or a computer." Smells like a bogus stat. Not that I'm arguing with the observation. But funny numbers is giving us a bad name.
I agree that the human brain is a marvel at pattern recognition. The result of millions of years of discriminating between food and what will make me food.
The simple matter is people and trained brains take time and cost money. If you can lower the cost and implementation time then many will regard an increase in false positives as a fair trade off.
@uk visa ... also, we have to segregate those who are *selling* solutions from those who are *buying*. I think it's someone *selling* this odour detecting technology. Those who sell have different motives and interests from those who buy. And, those who buy have different motives and interests from those tho *use*, e.g. us.
(I'd hate to see this solution get used).
Oh, sure, "Terrorists with murder in mind, drug smugglers, or criminals on the run are likely to be very fearful of being discovered," because they're all just normal folks; none of them are psychopaths or sociopaths. GMAFB.
Smell that, Rabbit?
*sniff sniff*... fear.
I wonder if the fear smells have molecular 'colors'. Elevated fear is orange.
Ordinary fear is yellow -- strike that. Yellow should be reserved for the absence of fear smell molecules. ("They call me mellow yellow...")
When does the smell of apprehension become the smell of fear?
I certainly agree with the first (righteous) remark of BF Skinner.
Seems to me a drug smuggler could take some drugs and relax to beat the system. As could anyone else. A terrorist might not, as it would dull their senses, but I would suspect that many would be relaxed at that point (I always relax just before a big event, you have done all the preparation, nothing else you can do, so relax).
Where is Starbucks on this issue, seems somebody hopped on caffeine might trigger this system, and that could adversely affect airport coffee sales.
What is going to happen is cascading failures. Think about it: One false positive and the next people in line will start to fear the check a bit. Then the next false positive and the fear will get stronger. When the finally check everybody, everybody waiting will be afraid!
Possibly the only people still getting through in such a scenario are hardened bad guys that have trained for this.
It used to be:
You must be guilty; the fact you're running away proves it.
You must be guilty; the fact you are afraid proves it.
All the while, the ultra-orthodox, with their utmost convictions and belief in a 'supreme being' guiding their act, will probably pass flawlessly.
I want a serious bonus if I my next job includes a lot of travel...
One of the warning signs of suicide is when a person who has been depressed is suddenly cheerful: they may have made a decision on how to stop being depressed. Similarly, people (including mdb above) have reported being relatively calm once the decisions are made, the preparations finished, and it's not time to act yet.
In contrast to other criminals, I'd expect terrorists to show less fear than average in airports and boarding airplanes.
"Seems like yet another technology that will be swamped with false positives."
It seems to me that this is research work.
The hope is that this could be used as a feature among many others for identifying potential threats and raising flags for potential human investigation.
We know it will work, because all terrorists are cowards. Politicians tell us this again and again.
See, so the problem solves itself, all the people getting tagged are ok, and the ones getting through are terrorists.
Man this is a bad idea. Next up, DNA scanners to scan for the 'evil' gene...
What is a terrorist afraid of? If we're talking about a guy just about to blow himself up -- what, he's gonna get caught?
Oh noes -- the worst case scenario is "You live"!
Now, a knowing mule may be afraid -- if they're not professionals, carrying a condom full of heroin in your intestines would make most folks afraid.
But a professional criminal? If his fear/stress level before action goes up past what the rest of us normally feel dealing with riding a multi-ton flying machine controlled by someone else -- well, then, we have little to fear from him.
"Terrorists with murder in mind, drug smugglers, or criminals on the run are likely to be very fearful of being discovered."
Now just hold on here, How do drug smugglers and criminals factor in? I thought this is all about trying to identify _terrorists_. This is just another example of the mission creeping beyond the "terrorist" boundaries.
"... When the finally check everybody, everybody waiting will be afraid!"
Ahh... But maybe that is the whole master plan here, sort of a reverse psychology. In the end, the only ones not afraid will be the terrorists, so they will be easy to identify.
Of course - I am almost always afraid when going through one of those checkpoints (when I'm not mind-numbingly tired or irritated to the point of rage). In the same way I'm afraid of the police when being pulled over - it's because I do not trust the people given that level of authority to use it justly.
"...we have to segregate those who are *selling* solutions from those who are *buying*..."
Excellent comment - and one applicable across a broad spectrum of cost-benefit analysis. I think Bruce has pointed out that we can usually trust strangers who *we* approach, but that we should look out for those who approach us with an offer.
I'm terrified of flying, and only slightly less afraid of arbitrary power given to semi-trained people in the name of "security".
And I look TERRIBLE in an orange jumpsuit.
hmmmm ... dangers of noise. There is a general background level of anxiety in the world (at least the US). Say they plug the thing it detects EVERYONE's fear. All the needles flip over to red and wrap around their stops.
I hadn't given much thought to masking agents like drugs. And this supports the "righteous" observation. Hassan loaded his assasins up on hashish before sending them into combat. The Lords army uses cocaine. Now if the thing could be modified to smell people with a buzz on or who had consumed massive (prescription) depresents in the US general population well that'a about 1 in 5 (business class the type 1 error goes to 1 in 2).
"...we have to segregate those who are *selling* solutions from those who are *buying*..."
Ditto to allowing them that have a thing to sell to define the problem space.
What I'd like them to develop is a world class BS detector.
@BF Skinner: "What I'd like them to develop is a world class BS detector."
Those should be mandatory in all election debates and advertisements.
Imagine this scenario. I'm getting on a plain to go visit my distant mother who is in the hospital. Just an hour ago I got off the phone with her and she told me they had discovered a mass in her head; they are uncertain at this point what it is.
I'm AFRAID/nervous for my mother's health.
I don't think this will work either. If they do implement it hopefully they'll recognize there are many reasons someone may be afraid. Also, what of the calm, collected terrorist who is honestly prepared to die?
@Phillip: "I don't think this will work either. If they do implement it hopefully they'll recognize there are many reasons someone may be afraid. Also, what of the calm, collected terrorist who is honestly prepared to die?"
For the record, I think this fear sensor is a bad idea. But in all likelihood it is just one more thing that would get someone scrutinized, but I doubt it alone would get someone kicked off a plane.
They already have people looking for people who act suspicious, I think they are under the misconception that if they automate something it may remove the bias of the observer.
I was once held at a gate until my connecting flight left so they could call it a missed connection rather than an involuntary bumping (we sat for 15 minutes at our gate but couldn't exit our plane until our connection, sitting at the next gate, pulled away). When I found out that they did this while we were on the ground next door, watching it leave, and they didn't have another flight for 24 hours, I blew up. The next day, I was flagged as a potentially hostile passenger and for the only time in my life I was subject to extra screening, but they let me on the plane.
This fear thing is a bad idea, but I suspect it will be used to determine who to screen more, not who to boot from the plane.
Bruce, email to you is bouncing:
Failed to deliver to 'email@example.com'
mail loop: too many hops (too many 'Received:' header fields)
Reporting-MTA: dns; mailfront1.g2host.com
(x's mine, just in case you don't want to share that address)
Maybe this device will be widely deployed and officials will realize they should stop fearmongering so they can reduce false positives.
There are several news reports to indicate that terrorists on suicide missions are typically medicated/drugged to prevent anxiety attacks from overcoming their bomb delivery "duties" or to prevent pain/shock if they're going on a shooting spree.
I think you're much more likely to catch a married white collar guy going on a booty call to some woman he met over the internet than you are a terrorist.
There are far more harmless neurotics in the world than dangerous neurotics.
Stress response is not mind reading. False positives out the arse. Dumb idea.
The TSA insists that their highly-trained (human) Behavior Detection Officers can reliably spot a person displaying the unmistakable signs of anxiety specific to terrorists. BDOs can reliably spot terrorists in a large room full of people who are bristling with signs of benign anxiety caused by things like the fear of being arbitrarily hassled at the checkpoint. That may seem implausible to ordinary mortals, but to a BDO the terrorist stands out like a blinking red light.
Numerous studies and tests have incontrovertibly proved that the TSA's classified training can and does transform ordinary mortals into infallible terrorist-detectors. But all that data is necessarily classified for National Security reasons, so we'll have to accept their assertions on faith. The TSA is highly effective at protecting aviation from the terrorist threat. That's a fact, backed up by seven years of classified data.
As with the BDOs, the effectiveness of every new measure the TSA adds to enhance security is proven by extensive testing before it's deployed at checkpoints. The tests and their results are, of course, classified for National Security reasons. But if they say it's effective, we should believe it. For why would a government agency we trust to protect us from horrible threats have any reason to lie?
@ Ian Osmond,
"And I look TERRIBLE in an orange jumpsuit."
Hey what about those nice shiny accesories "Steel Bling" braclets etc.
24 hour a day entertainment (if you like Big Bird and the rest of the pupets) at the sort of levels that just make you weep with joy.
Oh and the beard is a fashion statment you just cannot say no to (because no razors allowed you might cut yourself out of restraint etc).
The solution is clearly to decrease fear by relying on children to do the screening. I'm thinking of a black jacketed kid on a pedistal like out of Equilibrium.
I think we should simply get a new pledge of allegence. Instead of all of those words, just a hearty middle finger and 'F--- you, terrorists!' I bet that attitude change would decrease terror faster than any fear sniffer!
The TSA's operating procedures neither include nor allow the term "false positive." Those don't happen. Rather, every time one of the layers "alarms" it's counted as a "successful interception."
It's true that the BDOs, the virtual strip search, and now the terrorist-sniffer have yet to catch any terrorists. That is stark testimony to the TSA's value as a deterrent to terrorists!
Thousands of successful interceptions attest to the effectiveness of the TSA's layered security strategy that continually interdicts drugs, cash, fake military jackets, and numerous other illegal or suspicious items. The numbers speak for themselves as incontrovertible proof that the TSA's layered security strategy will protect aviation from any terrorist that happens to walk through a checkpoint!
And since the majority of "alarms" end with travelers cheerfully going on their way after a brief chat with a friendly officer, it is indisputable that the TSA can provide highly effective protection from terrorist threats with minimal delay and intrusion to passengers. The traveling public should appreciate the TSA's continuing protection, and welcome any additional layers of security that will make that protection even more effective!
If you don't want your posts deleted then you can try forming sensible arguments as to what exactly you disagree with, instead of immediatly resorting to namecalling.
I think there have been some perfectly reasonable arguments put forward as to why this will never be an effective tool, only serving to inconveinance the general population. Fear is a very common emotion with no direct connection to the intent to commit a terrorist act. To think that we can pick out terrorists by detecting fear is niave at best.
I don't think we'll ever see a culture that recognizes the importance of well-trained, educated, well-paid, experienced security and intelligence workers-- at least not in the US.
Anyway, I wonder if they could reverse the "polarity" on this test and find the 1 guy in line for acrowded flight who has zero signs of fear. That's where I'd want to start asking questions.
"The technology relies on recognising a pheromone - or scent signal - produced in sweat when a person is scared."
Let us just suppose the detector does work.
How long before somebody takes the idea and retunes the detector for other "pheromones"
Oh and by the way IIRC the glands that produce our "personal" smells are not those that produce "sweat" and are located somewhat lower than your arm pits...
(I would google to check but I'm not sure what other results would get pulled up and rumor has it in the UK the authorities look for "dodgy searches")
Dropping the whole terrorism pretense, I think it'd be really interesting to see how a technology like this might play out at the entrance to a sports stadium. On the one hand, I can easily see catching a whole bunch more people when they try to smuggle their own booze in (and increasing the revenue generated from selling $10 beers). On the other hand, I can easily see singling out a lot of businessmen who are attending the game with their clients, bosses, etc., and are anxious about any dealing or politicking that they might end up doing during intermission.
Would the beer revenue from smugglers be bigger, or the ticket revenue from anxious businessmen?
I imagine such a system would be nearly completely worthless as it'd also probably pick up every single person that had a drink before coming to the game, which I suspect, would be a very large percentage of the people.
Drug smugglers will just learn to take some fucking drugs before going to the airport so they aren't scared.
To answer an issue raised in the comments: Yes, humans are very very good at detecting threats and deception in other people. That's how our ancestors stayed alive. Unfortunately, our threat detectors rely on all kinds of stereotypes and prejudices that, while a useful heuristic in some circumstances, pose a problem for official security purposes.
The whole idea seems self-defeating:
To prove that it works, we'd need to test it on a significant number of "Terrorists with murder in mind, drug smugglers, or criminals."
If we can identify these people, we don't need the machine.
Of course, they've thought of that... we'll just assume it works and conduct the test in the field!
Terrorists are not generally psychopaths or sociopaths.
However, one *would* expect them to be experiencing less fear. In general, a person who has been planning to die, has gotten all the details worked out, and is finally about to do the deed is likely to appear relaxed and happy. (Read about suicides in your local paper, and like as not someone will be quoted as saying, "But it was such a surprise; he seemed to be on an upswing recently!")
As I believe I've mentioned before, I know someone who used to be a border guard on the Canada-U.S. border. His comment at one point was that you EXPECT people to be nervous trying to cross the border and facing customs guards, so nervousness is normal.
Instead, there are two classes of people you want to watch more closely: the really sweaty-palms nervous types who are almost certainly hiding something, and the smug 'I've got the perfect hiding spot and you idiots will never find it' types.
Every time travel to visit my parents and back I am terrified of something. I can't even explain what of. No, I don't think I'm afraid of flying. I guess it's rather having to deal with stubborn government officials who try to protect be (from what? from myself?). And guess what... because of that I sweat like crazy.
Why not train a dog to do this? Far cheaper, and more accurate, if it can be done at all.
And it'd make a good test of the concept: if a dog can't be trained to do it, it's a good bet that an artificial device can't be built to do this any time in the next few years either...
@uk visa "Why are we continually trying to find a technological way ... when we know that reading humans ... is something that intelligent well trained humans do better ...?"
Because our politicos claim that's discrimination or profiling. Computers and technology, on the other hand, are completely immune from profiling ... or so our politicos believe.
If our rules don't make sense, look at the rule-makers, and you'll see why.
Its already been said by other posters - fear (or the pheromone produced by it) doesn't constitute wrong doing. More garbage methodologies.
I'm also pretty sure someone can be trained to invoke a fear response in certain situations so this thing could certainly be underminded.
Jgreco @11:01 AM, that's good advice, but too late for "MaharishiSwamiPoobah", who has already been banned for exactly the kind of behavior he's displaying in this thread (plus repeated sockpuppetry). If he shows up again, please just ignore him -- his comments won't be around long.
Smells funny to me. Completely ridiculous. Who funds this stuff? I certainly can take a guess:
The taxpayers should take their money back. Even presuming this approach worked at all, it would generate more false positives that real results. What about people afraid of flying? What about nervous performers? How different are these from other stress pheromones?
So, how long before someone bottles the "fear" scent and starts spraying it around the airport? Are these security specialists really that stupid? I want my tax money back.
Better idea: maybe we can make a "TSA agent trap". It will work just like a Japanese Beetle trap, only instead of Horny Beetle Babe scent, we'll replace the lure with "Scared Human" scent. The agents will drop into a bag; after collecting a hundred or so, tie the bag and throw it away.
Ooooh, I gotta get working on the patent application.
I have a friend who hates to fly because he is always scared of the plane crashing. Surely this would give him another reason to hate to fly: being hassled as a terrorist because he is afraid of a plane crash (possibly even caused by a real terrorist). The whole idea is inventive, but ultimately stupid.
Rich, I just got a reply back after e-mailing firstname.lastname@example.org, so it's definitely working. If you're using an older address, please switch to that one.
Also, a couple of people reported error messages when posting comments during the last 12 hours; that should be fixed now.
Great, a self fulfilling prophecy: The paranoid fears the police hassling him, the police hassle him because he fears them.
All we need is one terrorist who is confident to shoot this one down.
My mom had a string of trips about 10 years ago (just *before* 9/11) where airport security took more interest in her, for no explicable reason beyond that she seemed a little nervous in line.
Fortunately, it only lasted a little while.
But for a while, she got *more* nervous any time she went to the airport, which seemed to turn into the same sort of "self-fulfilling prophecy" being described here.
Why do you folks assume this is some all-powerful technology? Why do you assume it works?
I expect this to be something along the lines of a breathalyzer. Everyone just assumes those things are perfectly reliable. But the reality is they merely detect molecules with methyl groups through spectral absorption. Ethanol is but one of a great many compounds that can trip the sensor. Your body naturally produces a number of such chemicals. (Eat well before testing, or your liver can go into Ketosis. (Ketosis breaks down fat into energy, in the process producing (releasing) methylated compounds.) Don't burp before testing! Don't breath car exhaust fumes at a road-side test before testing! Hope you didn't just paint your house earlier that day!) To say nothing of intentional failures using RFI (key-up a transmitter) to spoof unshielded electronics.
So now we've got this new technology. Managers and Directors being taken out to high-level dinners at 4-star restaurants on Caribbean junkets. It's sold. It's bought. Millions in taxpayers dollars go to line their pockets.
Citizens loose a little bit more of their freedom. False positive cases abound.
And terrorists slather on a few cents of perfume/cologne from walmart to defeat the system.
Looks like business as usual in the o'l U.S. of A.
What about the people who listen to the TSA and friends and are as afraid they want us to be?
I get it now... this doesn't only detect terrorists, but _potential_ terrorists!
After all, fear leads to anger, anger leads to hate, and hate leads to the dark side!
This terrorist boondoggle will run and run and run, until the politicians have taken away all our freedom and most of our money "to fight terrorism".
How many people were killed by terrorists in the US and Europe last year? ... compared with other sources of accidental death, like lightning strikes, kitchen accidents, etc...?
If one is afraid of being detected as a false positive, he'll be detected as having fear, so he won't be a false positive, actually.
Therefore you need to trust the system, or the system will catch you!!!!
Is this Sci-Fi or the latest-police-state-dream ?
Judge: "Accused, why did you produce that pheromon?"
Based on a pheromon, you get disadvantages; wouldn't that allow you to sue the TSA? Don't they need a proof to handicap people?
How does this square with President Kennedy's infamous "there is nothing to fear but fear itself" statement?
While they may have a fear of getting caught, the people this process is meant to detect would probably not be exhibiting the type or levels of fear necessary for detection. If they were THAT afraid, they probably wouldn't be carrying out these acts to begin with!
Beats me, why they keep trying to invent machines for what people do best... have a very fine antenna for what is going on at the subtle level.
Ancient vedic scriptures referred to this as the third eye.
This device sounds a lot like the Iraq bomb detection device (see http://www.nytimes.com/2009/11/04/world/... ). Utter nonsense, in other words. Snake oil, period. The notion that terrorists are any more scared than anybody else (especially with all the headlines about bad airline pilots lately) is a notion that, as far as I know, is utterly unsupported by any scientific data. Someone who believes Allah is going to give him 100 virgins in heaven for blowing up an airliner isn't scared, he's exultant, maybe, eager, but scared? Probably not.
Of course, given the fact that we have only had a handful of terrorist attacks in the past two decades and that the people who perpetrated them are by and large dead, I do suppose it's hard to conduct a scientific study. But this sounds more like a scam based on pseudo-science than anything else.
People seem to be ignoring some important points.
Firstly the research in question is still at a very early stage, being at the beginning of an 18 month feasibility study. Alot of such research leads to dead ends, which is the nature of research in general.
The fact that the article got the name of the lead investigator wrong a number of times makes me question any of the quotes atributed to the researchers, or any claims either for that matter.
The correct name is Professor Tong Sun (see http://www.city.ac.uk/sems/our%20staff/eeie/... ) who has been one of the leaders in the field of optical fibre based sensors for a number of years so if she thinks it is a viable proposition, I wouldn't be so quick to dismiss it.
"People seem to be ignoring some important points."
Yes most semed to have not thought it through a bit further.
As I said @11:07 on 3 Nov above,
"Let us just suppose the detector does work. How long before somebody takes the idea and retunes the detector for other pheromones".
Whilst I think sensing for "fear" is a wasted excersise due to too many false positives. A multiple pheromone sensor which also gives relative levels could easily form the basis of an "artificial blood hound" nose.
If each persons "personal smell" is held on a DB then in theory you would have a non invasive non contact bio-metric sensor.
Which might with a little work be able to sense an individual on a watch list at doorways etc.
The idea of the DB is not new certain "security forces" in the old communist block, reportedly stole peoples underware and stored in air tight containers to be used with real blood hounds.
The advantage of an electronic version would be that the DB record would not degrade and be almost instantly available.
As for potential it might be more usefull than a DNA DB for identification as DNA testing is currently impossibly slow and expensive for routien identification.
And yes the idea scares the heck out of me...
Perhaps the fear smelled by the TSA is the fear of being randomly selected for a body cavity search.
hmm... so then there will now be a technical justication to search anyone who "looks foreign" in an American airport? Because after cases like Maher Arar, I'd certainly be afraid if I didn't look like an unremarkable white middle-to-upper class businessman...
Schneier.com is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc.