Remember my rebuttal of Sam Harris’s essay advocating the profiling of Muslims at airports? That wasn’t the end of it. Harris and I conducted a back-and-forth e-mail discussion, the results of which are here. At 14,000+ words, I only recommend it for the most stalwart of readers.
ATN • May 28, 2012 8:13 AM
I believe the reason there is no major security check when you enter a train, a hotel or a shoping mall (potential targets for would be bombers) is that your are to be a costumer of someone who actually do care how you are treated when you enter the premises, and who would protect his costumers against obvious security abuses.
We can no more say that of air travel companies…
Nick • May 28, 2012 8:14 AM
Just finished all 14,000 wonderful words. One thing that struck me. Behavior and appearance profiling seem to be missing the point for me. Isn’t the thing we’re profiling to find actually just explosives hidden in ones crotch or other orifice? i presume baggage scanning is to look for the explosives in the carryon, metal detection for knives or guns. Seems like what we need is a better, faster and more efficient chemically assessing passengers. I recall the GE system that blew a puff of air was pretty ineffective, but I’d prefer we all got swabbed and chemically sniffed than irradiated, groped, etc. In 10 years, it doesn’t feel like the explosives detection aspect of the screening has evolved at all. In contrast, my laptop has been explosive screened 100% of the times I’ve been through Tel Aviv, with 0 false positives observed on my part and the travellers in line near me that I’ve been able to observe. it doesn’t seem like a stretch to scalably apply that same screening to ones person.
Dr. I. Needtob Athe • May 28, 2012 8:30 AM
When I read your title, for a second I thought you were swearing off talking about ethnic profiling at airports. 😉
Brian Utterback • May 28, 2012 8:46 AM
When I got to the line about no more readers at 13,000 words, I wanted to yell, “Yes there are!”
A wonderful format for this discussion.
Sam Greenfield • May 28, 2012 9:18 AM
I could only get through around 2/3 of the back-and forth. If I didn’t know otherwise, I would have thought you invented Sam Harris as a straw man. He consistently did not address your responses and kept trying to derail the discussion with irrelevant tangents.
Surjeet • May 28, 2012 9:28 AM
Excellent rebuttal Bruce! In fact, it would seem that profiling can actually make airport security much worse. Assuming that the profile is built on someone looking “muslim”, here is a simple terrorist plot: put two terrorists on the security line. The first guy “looks” like a jihadi terrorist – beard, islamic robes, Quran in hand, whatnot. His screening shows multiple anomalies. So he gets screened – but he actually doesnt have anything dangerous on him. But right behind him, is the second terrorist, dressed as say a fashionista with a “free tibet” t shirt. His screening also sets off anomalies – but if we are profiling and the security officers just had an obvious pick, it is very possible that the second terrorist would just be let through, and he can carry the real stuff. And in fact, for such a plot to succeed, we only need an order one fraction of security officers to fall for the trick.
andrew • May 28, 2012 9:34 AM
Interesting read. It did seem as if you repeated yourself many times to no effect. I still say any terrorist organization with an ounce of intelligence will do something different, negating the whole result of airport security. The security line itself is a far more likely target, since by definition it’s before security happens and also is a nice concentrated target. The assumption by the TSA and others that focusing on the last target will improve overall security is laughable.
Ulf Lorenz • May 28, 2012 9:46 AM
It is quite sobering to see how Sam Harris consistently fails to think his proposal through.
However, from reading your exchange, I would also suggest that you should have nailed him down much more. Give a broad overview about what you will discuss at which point, make him agree, and then block all his attempts to stray away from the topic at hand by either saying “this is irrelevant”, or “let’s discuss this one point later”. This is pretty harsh, but seems to me the only way to avoid his goal post shifting and whatever tactics he employs. You did it a bit, but seemingly not strict enough.
I do not believe that I would have succeeded to follow this recipe had I been involved in the discussion. It is just that I could not help noticing the similarity to discussions with climate change deniers or whatever.
Andrew Condon • May 28, 2012 10:05 AM
Thanks for going the full twelve rounds on this, Bruce. I think it’s potentially very useful to both non-security folks who get walked thru your approach at length and to security folks and psychologists who can marvel at how gut feeling can consistently throttle understanding in the mind of an intelligent person.
You might think it crazy but i could see this being given the “story of stuff” animated treatment….
Cory • May 28, 2012 10:11 AM
I’m an atheist, and a big fan of Sam Harris. I have a background in political science and criminal justice, a marriage of which is seen in this topic. I think you’ve thoroughly ousted Harris on this topic. I will happily support “wasting the $1” on current security inefficiencies after reading your cost benefit analysis. I particularly enjoyed your short essay “Refuse to be Terrorized” – I wish that I could see that philosophy embraced by our leaders.
Jim Lippard • May 28, 2012 10:38 AM
A cost of the sort of profiling Harris suggests which you allude to near the end but is never made explicit is that profiling which discriminates against Muslims carries a very real risk of causing radicalization. Scott Atran’s excellent book, Talking to the Enemy, points out that one of the factors that radicalized the Hamburg faction of the 9/11 terrorists was the experience of anti-Muslim discrimination. That helped turn secular engineers into Muslim radicals, and drove them to contact al Qaeda with their plans.
Harris suggests that Muslim garb might be useful in identifying Muslims to be screened–but doesn’t seem to recognize that the subset of radical Muslim terrorists who would wear Muslim garb while trying to get through security screening to perform an attack is already very close to zero, and if it’s part of profiling procedures, it’s definitely zero. If anything, a system that flags that someone who had previously been wearing Muslim garb but has recently shaved off his beard and dressed in Western clothes as worthy of further screening would make more sense than using Muslim garb as a sign for additional screening (see Atran: http://www.terrorismanalysts.com/pt/index.php/pot/article/view/35/html).
Erik • May 28, 2012 10:47 AM
This is pretty harsh, but seems to me the only way to avoid his goal post shifting and whatever tactics he employs.
Thing is, Sam tried to do that. At the beginning of the post, he tried to lay down guidelines saying they were talking specifically about airport checkpoints and specifically about suicide bombers. Then later, Sam didn’t like it when Bruce tried to stick to those rules, saying that you can’t talk about airport security in a vacuum and wanting to discuss mosques and intelligence gathering.
He was incapable of making his arguments work in the context he proposed, so he had to go outside of that context. Pretty sad, really.
Bob Duckles • May 28, 2012 10:56 AM
I read it all. I also follow and respect a lot of what Sam Harris writes. Bruce wins this one. Sam seems to loose it a little when it comes to Islam. Defining us vs. them is a big problem and declaring war gets us into a lot of trouble. The activities of terrorists are crimes against humanity. It would be better to treat them as crimes.
Beej Jorgensen • May 28, 2012 11:14 AM
“We saw this after 9/11, in the assaults and discriminations against innocent Americans who happened to be Muslim.”
Or, indeed, happened to look Muslim.
[Tech note: your link to essays-airline.html is missing an open quote on the href.]
Dena Shunra • May 28, 2012 11:24 AM
Harris’s admiration for Israel’s security is perplexing: yes, they have “behavioral profiling” but they also have a police/military state where everyone is either part of the military or under suspicion.
That (a) doesn’t scale, and (b) is not a desirable outcome for a society of equal citizens, and (c) is unsustainable without huge outside resources (much of which go to Israel from the U.S.).
I’d think that a security system that is desirable in the U.S. (or worldwide) should scale, live up to the ideal of equality under the law, and not require outside support coming to at least 5% of the national budget.
NobodySpecial • May 28, 2012 11:29 AM
Actually this could be extended to other aspects of airport security.
We know that drug users have long hair and look scruffy – so there is no need to check anyone else arriving from Columbia for drugs.
Fraud is committed by white guys in expensive suits so everyone else should be able to bypass customs.
nonegiven • May 28, 2012 11:29 AM
w/e.
I just know the terrorists have already won. I can understand metal detectors, swabbing everyone for explosive residue, x-raying the bags, searching all the people who test positive. The nudie scanners and the random searches have to go. The next terrorist incident is most likely to be in the security lines at the airport, none of this stuff will be effective against that.
“Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety..”
–Franklin
James Duval • May 28, 2012 11:57 AM
Wow. Almost painful, because Sam Harris didn’t seem to realise that anecdotal evidence and backing shaky opinions with more shaky opinions are not a counter-argument, they’re an admission of defeat.
We know that drug users have long hair and look scruffy – so there is no need to check anyone else arriving from Columbia for drugs.
Fraud is committed by white guys in expensive suits so everyone else should be able to bypass customs.
Haha, NobodySpecial I declare you misnamed.
Topher England • May 28, 2012 12:16 PM
I like the argument that “we should anti-profile — paying less attention to people who, based on the totality of their characteristics, could not conceivably be jihadists” because a main part of the to and fro argument is about error vs success rates. If terrorists try to recruit Ralph Lauren models, what success rate do they need to obviate Sam Harris’ model? Just greater than 0.
Brandioch Conner • May 28, 2012 12:34 PM
@Topher England
If terrorists try to recruit Ralph Lauren models, what success rate do they need to obviate Sam Harris’ model? Just greater than 0.
Why even bother to recruit? As Bruce noted, a bottle of blond hair dye is just as effective.
I’d go further and say that Sam Harris’ model could be completely invalidated by a group of Muslim Arabs simply dressing up as stereotypical Jews. Some faked passports and Mohammed al Azif becomes Isaac Goldstein and since everyone knows that Jews aren’t suicide bombers …
Still fighting my way through the 14,000 words. It’s difficult because Sam Harris keeps changing his position.
Dena Shunra • May 28, 2012 1:31 PM
@Brandioch Conner
interestingly, Goldstein (Baruch, not Isaac) is the name of a Jewish terrorist.
He moved to the Occupied West Bank from New York, and in 1994 attacked a prayer service in the Ibrahimi Mosque, killing 25 and wounding 129 Muslims in mid-prayer.
He has a wikipedia entry and fans, as well as some Israeli governmental support.
Of note: the first suicide bombing by a Muslim group directed at Israelis was specifically stated as being revenge for Goldstein massacre.
It would be inappropriate to profile all Jews (or all New York Jews, or all Hebron Jews) as terrorists. Same goes for all members of any other demographic.
Brandioch Conner • May 28, 2012 1:59 PM
That was a long read.
I think that you’ve finally met the anti-Schneier. He doesn’t think of movie-plot threats. He thinks in TV-show logic. Specifically “Star Trek”.
Race X uses race X weapons.
Race X uses race X tactics.
Race X members all behave the same.
Race X can ALWAYS be detected (even when disguised) by sufficiently advanced technology.
And by the same “logic”, race A will NEVER use race X weapons / tactics / whatever. Race A will ALWAYS use race A’s weapons / tactics / whatever.
So, since we can all agree that the facts show that Muslim suicide bombers used planes for the WTC attack on 9-11 we can conclude that:
Next, understand that all the people of the world can easily be sorted based upon “Star Trek logic” and you have an almost perfect profiling system. One in which you never have to worry whether sexy blonde new-age California woman is carrying a bomb to destroy a plane. Even if she were controlled by a Muslim to carry a bomb her behavior would show that a Muslim was controlling her.
Wael • May 28, 2012 2:46 PM
SH is misguided. I had a comment on several of his paragraphs with a counter example from past history – but chose not to post them. Ever heard of some “terrorists” disguising as rabbais to carry out an attack?
Bruce should have tabulated his cost / benefit analysis, or treated the problem as a pure security issue and showed SH a step by step threat analysis instead of letting SH jump all over the spectrum with “divergences”. The problem I see is that BS and SH disagree on the axioms, which may have contributed to the path BS adopted.
Gerald Fnord • May 28, 2012 3:42 PM
0.) When Harris confused careful, targetted, intelligence which aatempts to divine intentions through statements and actions with a moment’s visual assessment, he lost definitively.
1.) I couldn’t stop remembering a Ken Macleod quote:
Ignore your feelings; trust the targetting computer
lyalc • May 28, 2012 3:43 PM
Elderly people in wheelchairs are arguably more likely to be “terrorist minded” if being suicidal and disaffected are 2 criteria of terrorists.
Older, wheelchair-bound people are approaching a ‘natural causes’ death, may have serious illnesses that reduce their expectation of life duration, and may well be disaffected by any number of things that have happened over a long life.
As some of this population have less to lose, why not ‘go terrorist’ and make the world remember them?
aikimark • May 28, 2012 3:43 PM
Wow. Just finished reading and I thank both you and Sam for the wonderful exchange…civil to the end.
MauroS • May 28, 2012 4:05 PM
I like Sam Harris when he argues about religion, but here he was totally out of his field of expertise. He was all over the place and never really touched the main argument at hand.
I suppose Richard Dawkins, another famous atheist, would have agreed with Bruce: in biology there are many examples of disguises that evolved exactly to cheat profilers.
I tried to post at Sam Harris’ blog, but his blog does not take comments.
Shoemaker • May 28, 2012 4:16 PM
Airport security is bad enough at it is. I don’t want to imagine what transgressions they’d get away with if it only happened to “those people.”
Wow. I’ve never been a fan of Sam Harris in the slightest but even I was staggered at the abject lack of logical or intellectual rigour.
The man calls himself an academic, but would fail a high school debate class on that showing.
Kudos Bruce for remaining polite and trying to have a reasonable, objective debate, it must be frustrating to painstakingly explain the concepts of cost benefit analyses, base rate fallacies and complex system design to be met each time with ‘hurr durr you just need to search all the people with beards, because they look like Muslims, and terrorists always have beards’, or whatever the hell his point actually is.
Ben • May 28, 2012 4:33 PM
After reading the whole exchange I can’t help but feel that Harris cares more about using airport security as a way to punish Muslims than he does about actual airport security. I’m glad he’s not in charge of the TSA.
Dirk Praet • May 28, 2012 5:21 PM
@ Brandioch Conner
One in which you never have to worry whether sexy blonde new-age California woman is carrying a bomb to destroy a plane
My stand on the issue is that any theory or strategy based on an “Ab uno disce omnes”-fallacy (from one, learn all) is flawed by definition.
Sam Harris should also read up on his SciFi if his logic is based on Star Trek and the like. As every geek knows, Caprica Six in Battlestar Galactica is not the stereotype terrorist you’d expect to be sabotaging colonial defenses. I bet RSH can give even better examples.
Jack • May 28, 2012 5:35 PM
I remember an interview with SH, before the PhD and UCLA, etc. He impressed me as very small-minded, racing to accumulate credentials because his arguments didn’t stand on their own.
foo • May 28, 2012 6:23 PM
Bruce said:
“2010, non-Muslim Andrew Joseph Stack, who looks as American as they come, flew a small airplane into a building in a suicidal terror attack.”
The millions of Muslims in America, of all races and ethnicities, don’t “look” all american ? That’s a weird point to argue, from your side. Secondly, all caucasians are Muslim (the word Caucasian was invented by european anthropologists to specifically refer to chechnya). Generally, muslims like Turks, Syrians, Lebanese, Albanian, Caucasian, etc, have blonde/red hair. Thirdly, the non-caucasian muslims (from indonesia, africa) etc, outnumber the caucasian ones. But many of these africans and asians are non-muslim as well and the non-muslim ones outnumber the muslim asians (but both are huge in absolute numbers)
So let’s assume racial profiling works. What should we profile for…what race is a Muslim exactly ?
Pete Austin • May 29, 2012 3:31 AM
Agree with Bruce on most matters, but I think Sam Harris won this argument. Don’t expect many commenters to agree, because this is Bruce’s blog.
In medicine, nobody seriously questions that doctors profile people to select those at highest risk – e.g women for breast cancer (though men get it too) – then do screening tests, then try a proper diagnosis. Treating everyone the same would be madness.
Bruce fails to explain why this doesn’t work in airport security – pointing out repeatedly that profiling and screening are not perfect, which of course they are not intended to be.
NA • May 29, 2012 3:51 AM
@foo:
So let’s assume racial profiling works. What should we profile for…what race is a Muslim exactly ?
Obviously… “non Aryan”. It’s quite clearly where a fair amount of people are going without wanting to spell it out.
NA • May 29, 2012 3:55 AM
@Pete:
the big difference is of course that illness won’t try to infect people who were historically not infected in an attempt to decrease the number of treated people.
Such a thing might happen as a result of selection, though, but you’ve got to look at probabilities, rate of change/mutation, infection vectors, rate of population motion, etc. You can’t just say “it works for medicine, so it is deemed to work for everything else that’s very different” without actually looking whether it does.
Pete Austin • May 29, 2012 5:10 AM
@NA. Disagree. You have correctly identified one way in which screening is not perfect, but that misses the point, because nobody on either side is claiming that it is.
Oslo • May 29, 2012 5:23 AM
If I didn’t know otherwise, I would have thought you invented Sam Harris as a straw man. He consistently did not address your responses and kept trying to derail the discussion with irrelevant tangents.
That’s how he operates. He has one great big axe to grind and grind it he shall. You’ll get almost the same answers no matter what you happen to discuss with him.
Ping-Che Chen • May 29, 2012 6:08 AM
@Pete:
What NA described is actually a very important point IMHO. Basically, diseases don’t generally actively trying to bypass screening tests, as evolution works extremely slowly. However, terrorists do actively trying to bypass any screening effort. They don’t need 100 years of evolution to adapt. This makes ethnic profiling much less effective compared to normal disease screening.
Therefore, “screening works well for doctors” does not imply “ethnic profiling works well for anti-terrorism.”
mashiara • May 29, 2012 6:55 AM
Also didn’t yet have time to read the whole thing but back to the screening issue:
Ping-Che Chen explained a very important distinction, then there’s the whole issue of base-rate fallacy and finally we come to the question cost/benefit.
Of course if you assign infinite value to preventing an attack then even the most insane measures will seem lika a good deal but that’s definitely cheating (also from where do you plan on getting the inifinite resources required?)
Back in the real world where we have finite resources we need to spend the resources so that we get the most positive net-effect.
Mammography is relatively inexpensive and it’s not done every month (granted I don’t fly every month but that’s in part because I hate many aspects of commercial air-travel [since I cannot afford to fly first or even business class, so “economy”, or steerage as I call it, it is for me…]), I have no idea how many breast cancer tests are done in a year but I’m sure if you divide that by number of air-travellers you will apporach zero so it does not scale.
Another issue is that the “screening for testing” where you test women and skip the men because the risk in men is so low is also no comparison, because then you would simply have to test every single traveller: The probability of a given traveller matching a profile (a semi-trained person can evaluate in less than a minute, remember those scaling and cost issues) being a terrorist on an attack run are vanishingly small. Unless of course the profile is “is currently attacking you or civilians in your vicinity”, which has very low false positive rate but does not really catch anyone planning on hijacking a plane.
tl;dr: Re-check your understanding on Bayes rule and the base-rate fallacy. Think about where we could use all those resources better, because making an attack impossible will have infinite cost.
Erich Schmidt • May 29, 2012 8:09 AM
@foo
“Generally, muslims like Turks, Syrians, Lebanese, Albanian, Caucasian, etc, have blonde/red hair.” Well, I can say at least that Turkish muslims do not generally have blonde/red hair. Geez, visit Istanbul.
Clive Robinson • May 29, 2012 8:23 AM
@ Erich,
Well, I can say at least that Turkish muslims do not generally have blonde/red hair. Geez, visit not generally have blonde/red hair. Geez, visit Istanbul
Being just about 2meters tall and having a red (fairly bushy) beard, I can concure that “red heads” (or natural blonds for that matter) are actualy so rare in Turky that the locals in non “resort” areas especialy the older ones are genuinly and quite deaply effected by them and make strange signs to ward of evil curses.
It’s the same for quite a few Mediterranean countries…
Pete Austin • May 29, 2012 8:41 AM
@Ping-Che Chen: I agree that terrorists will try to get around airport security measures. But this is nothing new; in fact it’s universal. Bad guys always try to get around security measures, and they’ve been doing this for thousands of years.
This factor makes all security measures a bit worse. It doesn’t automatically make them worthless. And nobody in this discussion is claiming that screening is perfect.
@mashiara The “base rate fallacy” is relevant to using a test for diagnosis. For example I think medical screening for breast cancer selects over 10x as many people without the disease, compared to those with it. I expect security screening would be very much worse. But this is only about separating the population into higher and lower-risk groups, nothing more. If that can be done cheaply enough, then it allows diagnostic resources to be targeted slightly more effectively, even if the false positives rate is huge.
This sort of thing is routine in other fields – for example car insurance rates are influenced by factors such as age and address, because even a weakly-informed decision is better than a guess.
aikimark • May 29, 2012 8:54 AM
I think Sam Harris was trying to take a pragmatic approach with the security threat in its present state. Bruce was taking a more nuanced approach with his computer security experience influencing his points.
I would like to see a debate between Bruce and anyone about the benefits of moving toward a less-secure model that involves the traveling public’s acceptance of the increased risk. The cost and time benefits that offset the risk increase would be the selling points for acceptance.
Peter A. • May 29, 2012 9:13 AM
@Pete Austin:
The difference between the contexts of security and medical screening is very simple yet crucial: The terrorists, criminals, spies etc. are intelligent beings while bacteria, virii, mutagens and other causes of medical conditions are not.
paul • May 29, 2012 9:39 AM
The medical analogy also fails because it has a level mismatch that decoys one’s thinking. People at risk of a disease aren’t analogous to individuals “at risk” of being terrorists, they’re more analogous to flights at risk of being targeted by terrorists.
We could probably find characteristics of such flights (yeah, right) and profile people headed for those flights (just as we examine more closely the cells of someone at risk for disease), but the politics would be unsustainable.
poke • May 29, 2012 9:50 AM
BS: But so what? You’ve proposed a correlation between being Muslim and being a terrorist. … One, it doesn’t matter that the correlation between Muslim and terrorist is a causal relationship. We’re taking about a detection system. You’re proposing that we can detect attribute A (terrorist) by using attribute B (Muslim). That’s what matters, not whether or not there’s a causal arrow or which direction it points. In using the word “correlation” I was giving you the benefit of the doubt; it’s a lower bar. … I want to stick with your ethnic profiling system.
The context for Harris’s views on racial profiling of Muslims is Harris’s toxic opinions on preemptive nuclear bombing of Muslims, torturing Muslims, being at war with Islam, and Islam being the world’s greatest threat, and so forth.
Harris’s own words, in context:
We are at war with Islam. It may not serve our immediate foreign policy objectives for our political leaders to openly acknowledge this fact, but it is unambiguously so. —The End of Faith, p. 109
there is no set of beliefs … more imperiling of the future than the beliefs of martyrdom or jihad in the Muslim world which are central to the doctrine of Islam. —http://www.youtube.com/watch?v=j0pxN0DuMAE
I am one of the few people I know of who has argued in print that torture may be an ethical necessity in our war on terror.
Given what many of us believe about the exigencies of our war on terrorism, the practice of torture, in certain circumstances, would seem to be not only permissible, but necessary. —The End of Faith, p. 199
There is little possibility of our having a cold war with an Islamist regime armed regime … What will we do if an Islamist regime, which grows dewy-eyed at the mere mention of paradise, ever acquires long-range nuclear weaponry? If history is any guide, we will not be sure about where the offending warheads are or what their state of readiness is, and so we will be unable to rely on targeted, conventional weapons to destroy them. In such a situation, the only thing likely to ensure our survival may be a nuclear first strike of our own. Needless to say, this would be an unthinkable crime—as it would kill tens of millions of innocent civilians in a single day—but it may be the only course of action available to us, given what Islamists believe.
Whenever confronted with the irrationality of his arguments, Harris has backed away from the controversy, but not his own words. Racial profiling will be no different. Though it is commendable that Harris gave Bruce an opportunity to respond at his own site, if Harris’s past history is any guide, Harris cannot be expected to change his mind no matter what Bruce’s arguments are.
What else can one conclude but that Harris is a willfully ignorant and hateful bigot?
Ozy • May 29, 2012 10:03 AM
Sam Harris: “Where are the Tibetan Buddhist suicide bombers?”
Harris is still on about the “no Buddhist suicide bombers”?!
https://www.youtube.com/watch?v=emiwacyap0Q
Harris is just going to ignore, cherry-pick, and torture the facts until they tell him what he already wants to hear.
Charlie • May 29, 2012 10:52 AM
Another problem with the medical analogy is that medical profiling and screening is performed by highly trained (read: expensive) doctors and personnel, compared to relatively untrained TSA workers. They are also expected to make judgement calls, turning the principal-agent problem on its head for this case (yes, you might argue about the risk of malpractice suit, but doctors and hospitals get insurance for that).
Dean • May 29, 2012 11:26 AM
I am not a security professional, but I have enjoyed reading Bruce’s stuff for a couple of years now.
As a non-professional the idea of profiling targets for enhanced screening seems like common sense. It seems intuitively obvious that the threat is primarily from Muslims. I’m sure if you polled the general population you’d find wide agreement. From that perspective I think Harris’s argument is defensible.
Where I am disappointeded in Harris, however, is that for someone who purports himself to be rational he is remarkably oblivious to rational argument in this case. As has been mentioned, he is rather obsessed about the threat of Islam (not entirely without justification) and I think it is causing a huge blind spot in his cognition.
I very much enjoyed reading this exchange, and there is no doubt in my mind that Bruce won this debate in convincing fashion. To cling to a belief that profiling is a legitimate answer in the face of facts and reason requires almost religious-like adherence to that belief. Kind of ironic for one of the most prominent atheists in the country.
Sommer Gentry • May 29, 2012 12:45 PM
Harris is lying to us and/or lying to himself when he says he isn’t worried about consequences for people identified as false positives. SH says, “What is a false positive in the context of airport security? It might be nothing more than asking a person a follow-up question or performing a hand inspection of his bag. We are not talking about imprisoning people who fit the profile at the airport. A concern about false positives only makes sense if paying closer attention to innocent Muslims has some truly terrible consequences.”
But why are we even having this conversation if not for the distinctly sick, wrong feeling it gives us to see tiny blonde girls and elderly citizens frisked like hardened criminals and posing nude for anonymous men? The reason SH wants us to profile has nothing to do with better security – he just wants to make sure false positives happen to people who look like “them” and not “us”. People aren’t screaming stories of outrage across the internet because they’re afraid a bad guy will slip through. We’re outraged because we know it’s wrong to repeatedly probe and image a woman’s genitalia as punishment for her having a metal plate in her shoulder.
Harris gives his position away! He is “not worried about the “false positive” of executing a secondary search on an innocent person who fits the profile”, but if a false positive means searching an elderly veteran or other person SH considers sympathetic, then the screening is “a performance of security theater so masochistic and absurd” that he “thought he was being punked by the TSA”. Masochistic – doesn’t using that word mean he thought the screening caused the victims pain? And make no mistake, yes, rubbing a victim’s breasts and labia or penis and testicles, forcing frail people to stand in stress positions, re-traumatizing rape survivors, demonstrating dominion and control over someone’s sex organs, requiring acts of submission and surrender to the government, are all actions that cause real harm.
SH wouldn’t be in this fight if he didn’t know it’s a real harm to fondle Betty White’s private parts against her will. He’s only saying what he’s saying because he thinks directing airport sexual assaults toward Muslim women isn’t as harmful as screeners assaulting Christian women.
Steph Brady • May 29, 2012 1:44 PM
Actually I read the debate several time and recognized the problem.
Bruce was treating countering terrorists as a technical problem that ideally enabled a technical solution.
Harris (correctly) recognized that countering-terrorism against aircraft is primarily a people problem. Some technology may mitigate the risk but experienced people who are trained to counter it are much more reliable then any technology.
Profiling works because terrorists, like criminals, are not randomly distributed across the population. They have very clear profiles that can be identified by a combination of demographics & behaviors. We have a fairly large number of airline & suicide terrorist incidents in which to build profiles against. Random screening of a population looking for threats that are not random is counterproductive.
Its also important to recognize that any profile uses a feedback mechanism. As new intelligence comes in, it is used to update & refine the profile. Thus, terrorists have to try & recruit outside the profile (often difficult & resource intensive) and know that it can be countered with good intelligence.
Overall, I have to say Harris won the debate but it was certainly interesting.
Eric • May 29, 2012 1:54 PM
I think Bruce was setting up a straw man here. I like Bruce’ arguments in general but in this instance it was annoying.
Sam was not arguing that the ONLY characteristic that airport security should look at, was ethnicity/religion. He was arguing that these were relevant data that should be used in profiling. He’s right.
Bruce didn’t want to discuss any other topic, and didn’t even want to discuss the topic of ethnic profiling in the context of a broader system.
I would sum up my argument against Bruce as follows: don’t let the perfect be the enemy of the good.
Skukkuk • May 29, 2012 2:09 PM
Sam Harris says:
I don’t mean to belabor the point, but I would recommend the following exercise to our readers: Go to a public place — a restaurant, coffee shop, shopping mall, or an airport — where you can unobtrusively watch people go about their business, and see how much you can know about them just by looking. Ask yourself, what are the chances that those 20-year-old girls in yoga pants, buying frappuccinos, are taxidermists? What about the guy in his 40s, deeply tanned as though he never goes indoors, with tattoos covering both arms — what are the chances he’s a cardiologist? If you do this, you will begin to feel that you know the answers to these questions.
Right after that he suggests that during training, airport screeners could be tested for the accuracy of their judgments, which I guess means that they’d be led into a room full of people and only those who could spot the taxidermist would get hired. I doubt whether that’s a good proxy for people’s ability to spot terrorists, but at least it sounds like fun. But it’s the exercise he proposes for readers that really amazes me.
What he’s suggesting is that people find out how much they can “know” about people at a glance by going out and making a lot of snap judgments about people, without any way of verifying whether they were right. Not that it would help much if they could check their answers — you’ll rarely go wrong by guessing that random people are not taxidermists, and anyway, if cardiologists needed to sneak past airport screeners they would spend time in a tanning salon and get some bad tattoos. But what he’s proposing isn’t even a bad test of intuition; it’s no test at all. He wants us to “feel” that our intuitions are correct by noticing that they are strong. It’s like a Stephen Colbert bit about thinking with your gut, except he’s serious.
If the word rationalism actually means anything, this exercise is not rationalist. It is the antithesis of rationalism. If Harris’s commitment to rationalism is sincere, and not just a flag of convenience, then he needs to think long and hard about how he managed to go so wrong here.
eigenperson • May 29, 2012 2:48 PM
As far as I can tell, Harris’s argument rests on the following assertion: That he (Harris) can identify at a glance certain people who cannot possibly be terrorists.
This assertion rests on no evidence at all. Harris has never tested his ability to identify non-terrorists.
But he is very, very sure he can do it!
This makes for an impressively bad argument.
boog • May 29, 2012 2:59 PM
@Pete Austin:
…I think Sam Harris won this argument.
Aroo?
It seems to me that Bruce’s analysis of Sam’s proposed profiling policy was pretty dead-on. He clearly explained how dividing travelers into two tiers of security creates another avenue of attack, and how it is costly and reduces security and makes us less safe. What more did he need to do exactly?
…pointing out repeatedly that profiling and screening are not perfect, which of course they are not intended to be.
I think that was in order to explain how such imperfections in a needlessly-complex profiling/screening security procedure lead to ineffective security. Which Bruce explained quite well IMHO.
Brandioch Conner • May 29, 2012 3:57 PM
@Steph Brady
“Bruce was treating countering terrorists as a technical problem that ideally enabled a technical solution.”
I did not get that from that exchange. Nor has the been Bruce’s claim in the past. Are you reading the correct exchange?
“Some technology may mitigate the risk but experienced people who are trained to counter it are much more reliable then any technology.”
Strange. The available evidence seems to indicate the exact opposite. People become less observant over time with extremely unlikely events.
Again, did you read the correct exchange?
“They have very clear profiles that can be identified by a combination of demographics & behaviors.”
You mean like being Arabic or Nigerian or Jamaican or American? No. That has been covered sufficiently. There is no profile that matches them.
“Random screening of a population looking for threats that are not random is counterproductive.”
Actually, the math shows that you are wrong on that as well.
“Thus, terrorists have to try & recruit outside the profile (often difficult & resource intensive) and know that it can be countered with good intelligence.”
So the terrorists fit A profile except that then they will attempt to recruit outside of A profile so A profile will be changed to include the profiles of B and C and D and E … eventually everyone is in the profile of possible terrorists.
“Overall, I have to say Harris won the debate but it was certainly interesting.”
Despite every single one of his points (and yours) being shown to be incorrect. That’s the problem with trying to explain security to non-security people. They operate off of what they “know” to be “true” in their “gut” without any regard for what the evidence shows.
AP • May 29, 2012 4:29 PM
Have we all forgotten about Oklahoma City already? We’ve got a lot of terrorists to worry about and some are white and American. Invest more in investigation and counter-terrorism.
mashiara • May 29, 2012 5:21 PM
@pete
But this is only about separating the population into higher and lower-risk groups, nothing more.
This is not “nothing more”, see below for cost factors.
If that can be done cheaply enough, then it allows diagnostic resources to be targeted slightly more effectively, even if the false positives rate is huge.
The separating can be done cheaply (if you don’t care much for political cost or false-positive rate), but the “improved screening” is not free any way you measure it and the more false positives the more expensive this theater gets.
A few things to measure when evaluating the cost:
radicalizing those on the fence (hurting innocents is a very good way to convince people that you are in fact evil, just as the not-so-crazy-anymore -lunatics have been shouting all the time).
indirect cost to flight operators: people are flying less because of 3.
indirect cost to goverment (people driving instead of flying increased fatalities [travelling by road is actually pretty dangerous])
Now if you improve the profiling process that you use to deem low vs high risk you add cost there and still since the actual risk is vanishingly low you will always have horrid false positive rate.
@dean
It seems intuitively obvious that the threat is primarily from Muslims. I’m sure if you polled the general population you’d find wide agreement. From that perspective I think Harris’s argument is defensible.
“seems intuitively obvious” is definitely not evidence of anything. polling the general population is the worst idea ever. “People are not smart. A person is smart. People are stupid, panicky animals.” just about sums it even without accounting for the fact that media makes money by selling stories (and never let the facts get in the way of a good one…) that are emotional and memorable.
Threat coming from Muslims “seems obvious” to you because it’s the last memorable story you heard, not because there is some actual hard data on the risks.
It’s not that all security measures are worthless but the cost/benefit calculation is totally lopsided since the benefit from draconian measures has been assigned way too high value. Partly because the costs are externalized and partly because especially the political cost of a succesfull attack is large to those in charge personally (getting re-elected is important m’kay).
Insurance companies are businesses and I can choose my provider even when the goverment mandates that I must have liability insurance if I want to drive on goverment roads. No such choice with the security-screening: the only choice is the false one of not flying at all (in many routes flying is the only economical option [going by any other mode would cost, at least when remembering that your time is not free, many times the cost of flying])
mashiara • May 29, 2012 5:24 PM
@AP No, not a lot of terrorists, very few in fact. but focusing on “islamic terrorists” (aka. brown people with beards) is indeed quite stupid, domestic terrorism needs also to be taken into account.
Thus your final point of investigation and response stands.
Dirk Praet • May 29, 2012 5:40 PM
@ Steph Brady
They have very clear profiles that can be identified by a combination of demographics & behaviors.
As suggested by another commenter, Sam Harris’s case for racial profiling is primarily fueled by his overall opinion on islam and a paranoid obsession with AQ. Its the same kind of reasoning that at some point in history made pseudo-sciences like phrenology and physiognomy popular.
It’s a good thing that the likes of Anders Breivik, Ted Kaczinsky and Timothy McVeigh are not into airport or airplane attacks. But what makes anyone believe there’s no other non-muslim folks capable of turning to airplanes over some serious grudge they have with the US ? Can anyone really dismiss the possibility that a no-good-with-computers Dutch Anonymous supporter at some point blows a fuse over, say, the upcoming Assange extradition or the treatment of Bradley Manning and decides to take stuff to the next level ? Or some Chinese Foxconn employees hijacking a plane to crash it into Apple headquarters as payback for the family members they’ve lost ?
The proposal for racial profiling of Arabs and muslims in terms of security is a purely retroactive, frozen-in-time reaction to a set of specific events. It offers zero protection against attacks from any other corner than that of a weakened and dispersed organisation that – as seen with the shoe and underwear bombers – never had many competent field agents in the first place.
It is downright silly to assume that racial profiling will make any meaningful contribution in stopping plane bombers. 9/11 was made possible by a massive intelligence failure on behalf of America’s TLA’s, not by the shear genius of a gang of misguided muslim extremists. So can we finally move on and learn the rational lessons instead of obsessing over killer men with beards to the point that a permanent state of fear and emergency is slowly taking over common sense and constitutional values ?
I bet the Founding Fathers would be spinning in their graves as centrifuges if they had any idea of what’s going on today in the country they founded.
Pat Cahalan • May 29, 2012 5:57 PM
In addition to everything already discussed, Sam is making a fundamental error in population size. I’m not talking about the terrorist population.
Israel has twenty-seven airports, of which only one has over 2 million passengers per year.
California alone has 9 airports with over 2 million passengers per year. SFO has more passengers per year than the entire nation of Israel, and it’s not even the biggest airport in the state (LAX serves about twice as many people as SFO).
Never mind the cost factor, which would be prohibitive. Think about the applicant pool.
How would you ever find enough good screeners to populate every screening station?
Narine • May 29, 2012 8:46 PM
While Bruce’s argument is great, I feel Sam Harris still makes a valid point.
If there is an elderly white lady with no doubts about her nationality (“Barbara Bush” that Sam quotes), or a clear Caucasian or African American vet on a wheel chair, or an elderly Mexican lady, or a mid-western couple with a brood of completely Americanized children, allow them to pass thru far more easily – e.g. no shoes off for them, no laptop check. There are always personnel at the front of screening unit who can make this call and they will not be delaying the line due to such inconsistent standards.
Can this system be gained – absolutely can! But the existing system and checks can themselves be subverted as Bruce has pointed out on numerous occasions. Those routes of subversion are more of a sure shot than the profiling laxity proposed here.
We can make the security line move faster (by the screening personnel using their judgement to waive such people through) without compromising the level of security we currently have. I disagree that the screening personnel have no incentive to do so. If I were a screener, I would use my discretion to allow some harmless people through without a hassle, even if it meant my misjudgment could prove fatal. This is not an exact system we are talking about – there is already heck of a lot of judgement happening in interpreting the scanned images. Haven’t we read of guns in carry on bags being missed at the same rate they were prior to the enhanced security procedures?
This just loosens it a little bit more, and acceptably, I would argue.
On the flip side, I am quite positive that someone who is openly Muslim and religious is already getting more than the normal share of attention at the check point. If a person sporting a beard and a Muslim prayer cap passes through security, he is far more likely to be part of the “random” extra screening check and “random” explosive swab check. Guaranteed if he has a thick accent or a foreign passport from certain countries. The ticketing agents mark the boarding cards when they issue them itself, and we all have seen it.
steph brady • May 30, 2012 8:27 AM
“As suggested by another commenter, Sam Harris’s case for racial profiling is primarily fueled by his overall opinion on islam and a paranoid obsession with AQ”
Actually the 9/11 attack was the direct result of this countries refusal to take AQ seriously in the 1990’s. AQ declared war on the US & engaged in a serious of escalating terrorist incidents. returning to that level of ignorance would be highly destructive.
“Its the same kind of reasoning that at some point in history made pseudo-sciences like phrenology and physiognomy popular.”
Actually profiling is not pseudo-science. Its a valuable tool for identifying high risk individuals. Both criminals & terrorists have profiles that set them apart from the general population. The challenge is training people to recognize them.
“It is downright silly to assume that racial profiling”
The term “racial profiling” is a complete misunderstanding of the proposed solution. We are talking about profiling based demographics and behavior.
“So can we finally move on and learn the rational lessons instead ..”
Actually the goal is to prevent future airlines terrorism by identifying high risk individuals while raising the bar for trying to engage in terrorism.
steph brady • May 30, 2012 8:32 AM
“@AP No, not a lot of terrorists, very few in fact. but focusing on “islamic terrorists” (aka. brown people with beards) is indeed quite stupid, domestic terrorism needs also to be taken into account.”
We should focus on those terrorist organizations that have a history of hijacking aircraft. While this was done by some Marxist groups in the 70’s & 80’s, right now Islamic terrorist groups that are the highest risk of this type of attack. Its foolish for the TSA to focus on groups that dont engage in this type of behavior.
steph brady • May 30, 2012 8:45 AM
“So the terrorists fit A profile except that then they will attempt to recruit outside of A profile so A profile will be changed to include the profiles of B and C and D and E … eventually everyone is in the profile of possible terrorists.”
Actually profiling is a feedback mechanism. You use your intelligence resources to continually refine your profiles and detect if the terrorists are trying to change the profile.
Also, people underestimate the difficulty of recruiting outside of the profile. It is fairly difficult to recruit terrorists who are willing to take their families on suicide missions even though it can reduce the likelyhood of being profiled. Its also very difficult to recruit people born & raised in a western culture to become suicide bombers. The western culture does not view killing oneself to murder innocent people as particularly heroic. Other cultures have less inhibitions. Also, women are much harder to recruit to be suicide bombers then men. They seem to have a much higher reluctance to murder innocent people for a cause.
The point is that breaking out of profiles is difficult and very costly for the bad guys. Good intellligence can also provide ways to keep the profiles current.
mashiara • May 30, 2012 10:25 AM
@steph brady Yes if you only care about defending aircraft, as I understand a major point here is that focusing on single type of target and narrow set of tactics is not cost-efficient.
The challenge is training people to recognize them.
Also just how fast can these profile matches be evaluated ? Especially if you do not have the time to interview people (the Israeli approach simply cannot scale)
and again, what the false-positive rate ie. how much we increase costs to everyone (see my points above) and for how much benefit ? Then compare spending equivalent resources in more general measures to lessen impact of attacks or general intelligence operations and see what benefit do we get from there.
It’s not a question if measure X can be helpfull, it’s a question whether it’s worth the trouble.
Brandioch Conner • May 30, 2012 10:43 AM
@steph brady
“Also, people underestimate the difficulty of recruiting outside of the profile.”
Well since both males and females have been suicide bombers, yes, it would be difficult to recruit outside of that profile.
The problem is that most of the rest of the population fits that same profile.
“Its also very difficult to recruit people born & raised in a western culture to become suicide bombers.”
Richard Reid?
Born in England?
The reason we have to take our shoes off?
“Also, women are much harder to recruit to be suicide bombers then men.”
Really? So the female suicide bombers out there are harder to recruit? But they’re still recruited. And they still go kill themselves and their victims.
What is this “harder” that you keep claiming when there are specific instances of them already existing?
And finally …
“The point is that breaking out of profiles is difficult and very costly for the bad guys. Good intellligence can also provide ways to keep the profiles current.”
So the profile works to identify the terrorists and it is very hard for terrorists to not fit the profile except when they do and when that happens the profile will be changed to include the new terrorists who will find it very hard to not fit the profile except when they do …
I’ll make it easier for you. There is no profile that fits all terrorists (that also does not include 10,000x more non-terrorists).
Figureitout • May 30, 2012 11:33 AM
By only profiling for “muslims” or “islamists” domestically, we leave out 72.4% of the population (whites), 12.6% of the population (blacks), 4.8% of the population (asians), 0.9% of the population (American indian/Alaska Native), and 0.2% of the population (Native hawaiian and other pacific islander); for a grand total of 90.9% of the population out of the profiling process.
–according to: http://2010.census.gov/2010census/data/index.php
There isn’t even a separate category for muslims, they are that much of a minority. Of course, I’m making a big assumption that the Census data is accurate, as the U.S. is big. It would take a little more digging to see who among those categories identify themselves as “islamist”.
As an aside, I wonder if Mr. SH can figure out what I do and my personality just by looking at me. I don’t like to brag about much at all, but I have a fairly good intuition; but I would not feel confident more or less “guessing” someone’s occupation or in an airport security screening process their “malicious intent”…
boog • May 30, 2012 11:58 AM
@steph brady:
The point is that breaking out of profiles is difficult and very costly for the bad guys.
And yet they still do it. And by profiling we miss it when they do. So I’m not sure how that point is relevant.
MauroS • May 30, 2012 4:33 PM
I think Sam Harris and some people here are making the argument that “profiling by looks only can work” and maybe “those stupid liberals are ruining our security”. “Profiling by looks” can work IMHO and that’s not the question, the problem is if it can work against well trained, well financed and well-motivated terrorists drawn from a large pool of people. It can’t.
An example: a night club bouncer does profile people successfully “by looks”- I assume, since they cost money yet they are all over the place. Let’s suppose he tries to keep troublemakers and people without much money out, for instance. He has some big advantages over TSA:
The bouncer has just to raise the bar somewhat on average and act as a small deterrent. We ask nothing less than perfection from the TSA. And the terrorists can easily game the system – including using little children and old ladies, knowingly to them or not.
The solution to the problem of “assaulting little children and old ladies” is not to assault anybody unless clues other than “profiling by the looks” point to a problem, such as the result of good investigative work or the father of the terrorist turning him in.
Steph Brady • May 30, 2012 4:49 PM
@ boog
And yet they still do it. And by profiling we miss it when they do.
No, profiling is a feedback based process where profiles are updated based on the latest intelligence information and successful & unsuccessful terrorist incidents.
Profiling isnt a new concept, its been used for years by law enforcement and intelligence services. I think now that the lay person has become aware of it then they suddely perceive is a new thing.
Steph Brady • May 30, 2012 4:56 PM
@ MauroS
There is no “profiling by looks”. What Harris is suggesting is that a person having specific attributes might indicate they belong to a group that poses an above average risk. If so, then they should undergo a reasonable level of increased scrutiny.
This is something that people do instinctively as a survival mechanism. Go to any sketchy part of town and see if how you perceive people is the same for everybody.. I am willing to bet no.
Steph Brady • May 30, 2012 5:00 PM
@ Figureitout
I have a fairly good intuition; but I would not feel confident more or less “guessing” someone’s occupation or in an airport security screening process their “malicious intent”…
This is why training is so critical.. Lose 80% of the TSA people and train the remaining 20% in profiling techniques.
mepex • May 30, 2012 5:15 PM
@Steph Brady
“No, profiling is a feedback based process where profiles are updated based on the latest intelligence information and successful & unsuccessful terrorist incidents.”
In a system with tens of thousands of screeners across thousands of airports? Can you imagine the process and training required to implement such a scheme? This is at least an order of magnitude more complicated than a single one-time-decided two-tiered system that BS discussed, with more pitfalls and complexity.
Even assuming that each feedback loop is beneficial, which of course it won’t be, there’s no way the benefits outweigh the costs.
Brandioch Conner • May 30, 2012 8:09 PM
@Steph Brady
“This is something that people do instinctively as a survival mechanism.”
And people get it wrong most of the time.
And when they do get it wrong, those false positives usually do not negatively affect them.
But with airline security the false positives mean a LOT of wasted money.
“Go to any sketchy part of town and see if how you perceive people is the same for everybody.”
Why?
Hang out downtown during rush hour and see if you can tell which men beat their wives.
Then point them out for the cops to arrest.
How many false positives do you think you’ll get?
How many before the cops bring you in for causing a problem?
How many false negatives?
mashiara • May 31, 2012 1:47 AM
I’ll spell it out since no-one else has.
Yes, law enforcement and intelligence agencies use profiling, sometimes with at least moderate success.
However the succesfull profiling is not built on a minute of observation by semi-trained person, it’s built on research of background and carefull analysis by experts, possibly interviews (again by experts).
It’s really starting to look like some people are just looking for a “politically correct” way to limit the “enhanced screening” to “them” (“us” being white westerners [never mind that reid guy, or the man behind the curtain for that matter]).
Truly random selection is a much better idea, it’s fair (and needs to be implemented in such way that it’s seen as fair, this is very important) and has no worse “false positive” rate and it will nullify any advantage the attacker would gain by avoiding the profile.
Of course the random selection is going to let a terrorist through at some point, same goes for your precious profile. It is not possible to have perfect security (or anything for that matter but that’s a bit philosophical).
We need to do proper cost/benefit calculations, the whole politicians’ personal angle (both “I mustn’t look weak on terrorism” and “I need to get some gravy for my buddies at PervScan Inc” varieties) needs to be eliminated (now it adds a huge skew to the benefit side while actually only benefiting at best a handfull of people and not in a way related to airline security…)
Dirk Praet • May 31, 2012 4:21 AM
@ Steph Brady
Actually the 9/11 attack was the direct result of this countries refusal to take AQ seriously in the 1990’s
If this is so, then you just made a most convincing argument for racial/demographic profiling being totally useless, i.e. “These bearded AQ dudes from Afghanistan and Saudi Arabia are way too retarded to pull off anything serious against us”. Not. A slightly different attitude, inter-agency collaboration and connecting the dots could have made all the difference.
Again: as discussed more than once on this blog, Israeli-style, intelligence and behavioral based profiling definitely has merits, but doesn’t scale rather well. What SH is proposing is something completely different.
Blackfist • May 31, 2012 7:48 AM
About female suicide bombers, there’s about 10 times more male than female suicide bombers to date. There is 125 confirmed acts of suicide bombing performed by women and there is a bit over 1000 acts of suicide bombing performed by men.
The efficacy in terms of wounded or deaths per attack is about the same for both male and female suicide bombers.
I looked this up on the CPOST database.
Steph Brady • May 31, 2012 9:45 AM
@ Brandioch Conner
You are missing the point. Even untrained people can recognize threats at a subconscious level. Read the book “The Gift of Fear” for more information on this topic.
But what we are talking about here is Training people to spot individuals that meet certain profiles consistent with people who engage in terrorism.
Steph Brady • May 31, 2012 9:50 AM
@ Dirk Praet
Actually, What SH is proposing is using a combination of demographic & behavioral characteristics to identify individuals who are members of high risk groups. This is what the Israeli’s do.
People just like to call this “racial profiling” which is inaccurate because its not actually about race.
Steph Brady • May 31, 2012 9:55 AM
@ mashiara
“Truly random selection is a much better idea, it’s fair”
If you are building security processes based on “fairness” then you are pretty much ensuring failure.
Both criminals & terrorists have specific characteristics. Ignoring this to focus on people who do not pose a risk ties up valuable resources while producing no significant increase in security.
boog • May 31, 2012 11:56 AM
@Steph Brady:
No, profiling is a feedback based process where profiles are updated…
So? You said that breaking out of profiles is difficult and costly – I pointed out that bad guys (specifically the strategizing kind) still do it, and by using profiling as part of a security system we are likely to overlook said bad guys. In what way does the definition of profiling preclude such a concept? If a bad guy wants through security badly enough and fitting a “good guy” profile will get him (or her) through with less scrutiny, what do you think he/she is going to do?
Profiling isnt a new concept…
In no way did I suggest that it is.
…its been used for years by law enforcement and intelligence services.
Right, but as a security filter for the general population?
Brandioch Conner • May 31, 2012 5:06 PM
@Steph Brady
“Even untrained people can recognize threats at a subconscious level.”
Then you should be able to, on a subconscious level, identify men who beat their wives.
If you cannot, then you cannot identify threats.
All you can do is identify that which makes you (personally) FEEL threatened.
Big difference there.
Lots of people would be afraid of a tarantula.
Yet the worst a tarantula can usually do is cause discomfort (not death).
Do not confuse your fears for a security system.
Steph Brady • May 31, 2012 5:19 PM
@ Brandioch Conner
“Then you should be able to, on a subconscious level, identify men who beat their wives”
You misunderstand the issue.. Our subconscious (gut instinct, intuition, many different names) is a survival mechanism that identifies high threat situations. With training, this can be used to identify people who pose a specific risk.
Now Profiling also works at the conscious level where you consciously look for people who have specific characteristics associated with a type of criminal or terrorist. The DOJ has lot of data on the variables associated with types of criminals.
To use both of these effectively, you need training & experience. This is where the detecting wife beating scenerio fails. The average person is not trained to do this.
Steph Brady • May 31, 2012 5:30 PM
@ boog
If a bad guy wants through security badly enough and fitting a “good guy” profile will get him (or her) through with less scrutiny, what do you think he/she is going to do?
Security has never been about keeping every bad guy out. Security is about making it as difficult as possible for the bad guys to achieve success by raising the cost. Trying to recruit elderly west european women traveling with children (completely out of profile) to become suicide bombers, without detection, is much more difficult then recruiting young disenfranchised arab males (based on # of suicide terrorist attacks).
Security is about trying to secure resources in the most cost effective manner. Profiling is a cheap way with a low cost for false positives (time spend doing secondary screening &/or interviewing person for more information)
Brandioch Conner • June 1, 2012 12:26 AM
@Steph Brady
“With training, this can be used to identify people who pose a specific risk.”
Such as men who beat their wives.
Except you say that it cannot.
Which was my point. It is not about a “threat”. This is about you FEELING threatened.
Bruce, this is why I would prefer different terms for real security as opposed to the feeling of being secure.
“Now Profiling also works at the conscious level where you consciously look for people who have specific characteristics associated with a type of criminal or terrorist.”
We’ve also been through that. Repeatedly.
The only characteristic you’ve been able to show is “looks scary (Muslim) to me”.
Which, again, is easy to defeat even if it was a valid characteristic (which it is not).
“To use both of these effectively, you need training & experience.”
Except that they are not effective as has been pointed out. Since they would not have stopped Richard Reid.
“This is where the detecting wife beating scenerio fails. The average person is not trained to do this.”
No.
That example was to illustrate how your claims failed.
You were the one who claimed (as you just did):
“Our subconscious (gut instinct, intuition, many different names) is a survival mechanism that identifies high threat situations. With training, this can be used to identify people who pose a specific risk.”
If that were accurate then you would be able to identify at least some of the wife beaters. Since you cannot, your claim fails.
Dean • June 1, 2012 10:46 AM
mashiara said in response to my comment:
“seems intuitively obvious” is definitely not evidence of anything. polling the general population is the worst idea ever. “People are not smart. A person is smart. People are stupid, panicky animals.” just about sums it even without accounting for the fact that media makes money by selling stories (and never let the facts get in the way of a good one…) that are emotional and memorable.
Threat coming from Muslims “seems obvious” to you because it’s the last memorable story you heard, not because there is some actual hard data on the risks.
While I appreciate your comment, perhaps you didn’t read mine in its entirety. If you had, you would realize that I offered that up as a reason people like Sam Harris would feel that profiling is a legitimate exercise. In fact most people, lacking security expertise (or not having read Bruce’s essays on the subject) would feel the same.
As you alluded, it is largely a result of the availability heuristic and its effect on risk perception. The cases of muslim terrorism are easy to recall, so we tend to both exaggerate the risk and, specifically, attribute the risk to muslims, or “muslim-looking” people.
I’ll repeat: no rational, thinking person could read through the Harris-Schneier debate and think that Harris actually won.
Steph Brady • June 1, 2012 12:40 PM
@ Brandioch Conner
We are not trying to profile wife beaters, we are trying to profile people who want to hijack Aircraft. Completely different issue & profile. Its best to focus on the problem at hand.
The only characteristic you’ve been able to show is “looks scary (Muslim) to me
Actually, this is incorrect. The FBI does terrorism profiling with their System to Assess Risk (STAR) system & their Behavioral Science Unit. The Israeli’s do this in their airports.
People may not be comfortable with the idea but that certainly does not mean its not effective.
Steph Brady • June 1, 2012 12:52 PM
@Dean
no rational, thinking person could read through the Harris-Schneier debate and think that Harris actually won
Actually Harris seemed to understand the problem better. He recognized that profiling, terrorism and airline security at its heart is a people problem. Its about identifying likely threats. Bruce tried to treat the issue more as a security engineering problem.
The problem was that neither side discussed what profiling actually meant. Their was this “looks Muslim” nonsense. They should have discussed profiling based on demographics & associated behavior. Unfortunately Bruce seeemd to fal more into the “looks Muslim” fallacy then Harris did.
Brandioch Conner • June 1, 2012 1:32 PM
@Steph Brady
“We are not trying to profile wife beaters, we are trying to profile people who want to hijack Aircraft.”
So you are saying that there is something about a hijacker that makes is possible to profile them but makes it impossible to profile a wife beater?
Could that something be …. Muslim?
If not, then please explain how the wife beaters evade profiling. And what is to prevent the hijackers from using the same techniques as the wife beaters.
“Completely different issue & profile.”
But if it works then it works. And if it works then it should be easy to demonstrate by going downtown during rush hour and identifying men who beat their wives.
“Actually, this is incorrect.”
No. It is correct. You keep referencing things that you find scary and then extrapolating to define those things as threats.
Like the tarantula example, scary looking is not a threat.
Like the wife beater example, not scary looking can be a threat.
“The Israeli’s do this in their airports.”
You should read Bruce’s website more often.
The Israeli method has been discussed and analyzed here many times.
First off, it does not scale. The entire nation of Israel moves fewer air travelers than a single airport in California.
Secondly, the Israelis use “behavioural profiling” rather than “does this person look Muslim”. This allows them to catch instances where the person is not even aware that they are carrying a bomb.
Thirdly, the Israelis start their security process outside of the airport, on the approach to the airport.
Lastly, as Israel demonstrates, when a terrorist sees that airport security is more difficult than blowing up a bus, the terrorist will switch to the easier target.
It is easy to claim that method X will stop Y when the odds against Y are millions to 1.
But if method X really worked then it should also work against A and/or B and/or C.
Domestic violence is a lot more common here than terrorist attacking planes. If that method worked then it should be able to identify men who beat their wives.
Or you should be able to explain exactly why it would not work.
boog • June 1, 2012 1:38 PM
@Steph Brady:
Trying to recruit [one extreme example] to become suicide bombers, without detection, is much more difficult then recruiting young disenfranchised arab males (based on # of suicide terrorist attacks).
I feel like we’re arguing in circles here – my point has been that in spite of the difficulty and cost, it still happens and when it does it’s a bad thing. Obviously security isn’t about catching all bad guys, but it also isn’t supposed to create new attack vectors. Consider that my objection isn’t over Sam’s idea to tighten security on people who fit the “bad guy” profile, but rather the implication that security would become relaxed for people who fit the “good guy” profile.
Security is about trying to secure resources in the most cost effective manner. Profiling is a cheap way with a low cost for false positives (time spend doing secondary screening &/or interviewing person for more information)
I’m not so sure profiling would be as cheap as you suggest, but even with a low cost per false positive, with so many false positives the costs would become quite large.
That said, I’d still be quite concerned about the false negatives.
Sushil K • June 6, 2012 1:43 AM
Nothing to do with airline security, but how Israeli authorities are implementing screening: “Show us your private email.”
Suhil K • June 6, 2012 1:46 AM
Sorry, URL fail. I was trying to link to a Gawker article about Palestinian-origin women who were subjected to seemingly ridiculous levels of scrutiny at the border.
Subscribe to comments on this entry
Sidebar photo of Bruce Schneier by Joe MacInnis.
gregorylent • May 28, 2012 7:20 AM
surprised people pay attention to sam harris … automatically disqualified for being a careerist academic