Thursday, November 14, 2013

Blame Detector, Not Behavioral Detection

As the Government Accountability Office calls TSA to task for catching no terrorists and realizing no verifiable security benefit from its behavioral detection program, the popular temptation is to demonize the tool instead of its ham-handed implementation. (For details see http://p.washingtontimes.com/news/2013/nov/13/tsa-wastes-money-profile-passenger-behavior-report/ )

That would be a mistake, the kind that perpetuates the myth of racism reflexively attached to the term behavioral profiling out of a rash equating of all profiling with racial profiling. Let us begin by clarifying terms in order to put the pin back into the grenade that pejorative labels have become.

Racial profiling is stereotyping at its worst, usually associated with authorities singling out minorities for invasive attention or arrest on the basis of their skin color instead of on the basis of probable cause. This is reprehensible and inexcusable -- as is any abridgment of constitutional rights or due process under any smokescreen offered to legitimize it.

Behavioral profiling is altogether something else. Its only relation to racial profiling is that both terms use the word "profiling," which is not enough to make them synonyms. Otherwise, progressive agenda would be indistinguishable from conservative agenda, financial asset would be the same as financial liability, and confidence builder would be no different from confidence man. After all, one word is the same in each pair of two-word labels. Please acknowledge the weak logic behind making such definitional leaps.

No, behavioral profiling owes its place in the quiver of security arrows to Israeli security screeners for El Al, who are to TSA what a surgeon is to a butcher. The signature case establishing the security value of this technique involved catching a pregnant Irish woman with a bomb who looked nothing at all like an Arab and who did not herself know that she was carrying Semtex concealed in her luggage onto her flight to Tel Aviv. What happened? Her Jordanian boyfriend targeted this woman as an unwitting agent, wooed her, got her pregnant -- all purposefully in order to guarantee that she would fit no traditional stereotype. Consequently, detecting her by "racial" profiling would have been impossible if El Al screeners were only looking for young Arab males who fit some preconceived list of what a Hollywood filmmaker would ask Central Casting to use in advertising for someone who looks like a terrorist. So, the terrorist was himself betting on racial profiling and ready to bypass it.

Now we see where the behavioral clues took over to unmask this plot. The essence of the behavioral technique involves asking questions to pierce through the kind of cover story that villains must use in order to get through security screening. Using this technique is more akin to counterintelligence than police work. It takes a thinking questioner to drill down to the point of spotting where the cover story breaks down. And this requires a supple mind rather than the rote grinding through of a checklist. Thus, the El Al screener asks the purpose of the traveler's flight and engages in conversation to validate that the answers make sense.

In the case of the pregnant woman who was unwittingly carrying a bomb, her story just did not wash. She was going to meet her fiance's family, but he was traveling by separate flight. She was going to be met by people she did not know and did not have enough money even for cab fare. In reality, she was in love with the boyfriend and father of her child and, as a result, was understandably gullible. Not so for the El Al screener, however. Spotting the inconsistencies in her story, he used the behavioral technique to flag this passenger for extra scrutiny. This scrutiny, in turn, found the Semtex before it made its way into the cabin to detonate in flight and take over 300 lives. (For details on this particular case, look up the 1986 Hindawi affair and the name Ann-Marie Murphy, the pregnant woman, and her paramour, Nezar Hindawi. A place to start is http://en.wikipedia.org/wiki/Hindawi_affair)

Bottom Line: Behavioral detection works, if properly implemented.

Two big hurdles for the Transportation Security Administration limit effective TSA use of this technique. First, the persistently negative popular association that clings to every appearance of the word "profiling" makes it almost impossible to discuss this matter without unleashing a torrent of diatribes against the evils of racial profiling. Even when true, these accusations are beside the point and an argument unrelated to security screening. A related problem, though, is that the masters of this technique have a fondness for the word "profiling" when describing and teaching what they insist on calling behavioral profiling. To its credit, TSA has rebranded this method as behavioral detection, but the old term survives and all the baggage of "profiling" taints serious discussion of the technique's value and proper application.

Second, TSA implementation of behavioral detection is what merits closer attention than a technique itself which has been proven in the crucible of aviation security. If a technique works but the people applying it don't, we must ask what is wrong with the larger picture.

Behavioral detection is like a medicine capable of curing an infection. It is not enough to prescribe the medicine. It is also necessary to administer it properly, to take it the right way at the right time. An analytical observer would do well to see how an El Al security screener applies behavioral detection and then compare a TSA screener's approach. The screeners may be the same age and test at similar IQ levels. However, they operate in different environments, under different expectations, and with different enabling or constricting circumstances. The Israelis cannot afford to make a mistake. They live under omnipresent, existential threats. They also operate with more responsibility and with bosses and customers who trust them with life and death decisions. What about their TSA counterparts? Reports to date suggest that TSA screeners operate at a much lower level of discretion, responsibility, and applied judgment. Behavioral detection requires more than just following a checklist, more than a go-to-the-freezer-and-get-the-box mentality that sets apart a chef from a warmer of TV dinners. Both screeners may ultimately come from a gene pool that is more similar than it is different, but their management and training are critical in distinguishing between success and failure when it comes to applying a useful technique.

The GAO indictment of behavioral detection is misplaced. It is not the technique that deserves to be questioned so much as the management and context of its implementation.

-- Nick Catrantzos