TSA security test results: Replace humans with machines

By now, everyone has read about the recent Department of Homeland Security’s Inspector General’s field tests to ascertain the effectiveness of TSA’s airport security checkpoints. TSA failed the tests miserably 67 of 70 times (96 percent).
Since the results were revealed, Homeland Security Secretary Jeh Johnson relieved TSA Administrator Melvin Carraway and issued a series of directives to “enhance our security capabilities and techniques.”
The directives include: revising checkpoint operating procedures, intensive training of Transportation Security Officers (TSOs), retesting and evaluating screening equipment, and ensuring equipment is properly maintained.
An analysis of TSA airport security overall points to a systemic problem, not operational issues, at the heart of the checkpoint failures. No amount of training can overcome basic human traits.
After passenger IDs are verified at TSA’s airport security checkpoints, passengers are required to have their carry-ons and personal belongings X-rayed. TSOs check for weapons, explosives and banned items via X-ray units with color enhanced displays. The color enhancement greatly assists TSOs in identifying what passengers are bringing through the checkpoints, then to their flights.
Unfortunately, human beings aren’t well equipped to screen passengers’ belongings via the X-ray monitors, even with enhanced displays. It has to do with how humans perform pattern recognition and matching. Humans are tremendously adept at picking out patterns from random data. It’s an important survival skill we’ve developed. For TSOs operating the X-ray units, the problem is that while humans are great at detecting patterns in random data, we are not very good at detecting exceptions in uniform data, which is precisely what TSOs looking for weapons and explosives are asked to do.
Unlike humans, well-programmed computers are adept at pattern matching and finding exceptions. Facial recognition software is a great example of a computer’s ability to pattern match, even when disguises are used. TSA needs to contract for development of checkpoint X-ray unit computer recognition software, programmed to “see” weapons and explosives in real time. Computers should be efficient in finding such items with few false positives.
While their belongings are being X-rayed, most air travelers are sent to be scanned in full-body scanners using MMW waves to check if they’re carrying weapons, explosives or other banned items on their bodies or in their clothing. The scanners cost about $141,000 each, and data shows they have serious shortcomings which any dedicated terrorist can easily exploit.
There’s much evidence from the TSA Blog that the scanners will catch travelers secreting guns, knives, martial arts stars, etc., on their body and in their clothes. Of course, so will $4,000 metal detectors. The scanners have a reasonable chance to detect a home-made plastic gun, or a ceramic-based gun, something metal detectors won’t find. In all likelihood, the scanners won’t detect low density explosives taped to an air traveler’s body, and they can’t detect weapons or explosives secreted in body cavities, since MMW waves can’t penetrate the skin.
Therefore, due to their shortcomings, expensive full body scanners will often fail to detect explosives carried by a terrorist who has any intelligence at all.
TSA needs to recognize that passenger screening will never be perfect, and it doesn’t have to be. Strategically, passenger screening only has to be reasonably good to stop terrorists. When there’s a better than even chance they’ll get caught, terrorists won’t try to pass through airport security checkpoints, but will use other means to access airplanes, or move on to other targets.
Moreover, knives and guns are much less of a problem than explosives, which can be more devastating to a flight, and more likely to go undetected by our current airport checkpoint systems. In part, what makes this true is what TSA calls its “twentieth security layer”: passengers. Since 9/11, passengers have shown they are prepared to take on terrorists in-flight, and have been successful subduing them.
TSA needs to dump the seriously flawed full-body MMW scanners, go back to magnetometers, and add low-tech, but effective, explosives sniffing dogs to the security mix.
Beyond these two major changes to TSA systems, better training and even raising the minimum hiring standards for TSOs couldn’t hurt, but if TSA is to really improve their effectiveness, they’ve got to overhaul their systems, which primarily depend on security checkpoints at US airports. Currently, of TSA’s $7.3B budget, $4.5B is spent on airport screening. That’s about 62 percent of their total budget, much of it spent looking for pocket knives, large shampoo bottles and toothpaste tubes, and even toy ray-guns carried by ordinary passengers, not terrorists.
TSA should be focusing on intelligence. They should be focusing on the terrorists when they are planning their attacks. They should be utilizing their resources to keep terrorists far from airports. Intelligence should comprise the vast majority of TSA’s budget.
When TSA confiscates a 5-ounce container of shampoo, they’ve failed because they spent their time hunting down a false alarm. When intelligence stops a terror plot like the British did in 2006, stopping terrorists who were planning to blow up seven airplanes with liquid explosives, that’s a success.