Hello, law enforcement. I suspect you’re reading this because, as a TSA supervisor told me recently, “… we are interested in you”.

Yes, I asked to fly selectee–to not provide ID–at Denver International recently. Yes, I’ve done this before. Yes, there was a lot of confusion between TSA employees on whether that was legal or not–eventually M. Gatling of the DIA police told me I was required to display ID. Yes, I opted out of AIT. Yes, it did take no fewer than eight TSA officers, airline representatives, and police about 45 minutes to determine I posed no threat. Yes, I was exceedingly polite, and most of us got along quite well. Yes, I was asked all kinds of questions I was under no obligation to answer (among them my address and phone number), and no, the TSA supervisor was not very pleased that I asked whether I was legally required to respond.

“What is your contact number.”

“Am I legally required to give that information?”

“I’m asking you.”

“Well, that much is clear.”

“What is your contact number.”

“Am I required to tell you?”

“I’m asking.”

…

To cut to the chase: no, I am not a terrorist. No, I have no interest in harming anyone. Quite the opposite, in fact. If you’re interested in why I disagree with the screening system, my reasoning is simple; the screening process is not sufficient to detect probable threats, yet incurs disproportionate social and monetary costs.

The costs are obvious: long wait times, missed flights, lost items, and depending on your personal views, dignity. Some people are not okay, for a variety of powerful reasons, with having their body touched everywhere. The question is, are these costs proportionate and acceptable given the process’ effect on the probability distribution of bad events–i.e., to what extent it prevents people from crashing planes into buildings.

I believe the answer is no. The TSA’s strategy for passenger screening has been primarily reactive, not anticipatory. The organization simply doesn’t think ahead to consider probable threats. Consider that we only started to remove shoes for screening after Richard Reid’s attempt to light his shoes on fire in 2001. Similarly, we only began limiting liquids after the 2006 UK bombing plot. Toner cartridges? Same story.

The problem is not necessarily that these systems fail to prevent the attacks they are designed to combat–it’s that they do not address other, previously unexploited avenues of attack. The specific failures in these systems include:

1. The pat-down is insufficiently aggressive. I’ve opted out of AIT several times and received their pat-down. They don’t reach between the buttocks, nor do they check behind the scrotum. I’m a relatively trim person. Someone with significant body fat could likely conceal a reasonably sized weapon in this area and avoid detection. The only explanation I can conceive of is that the TSA has modified their pat-down to prevent public exasperation. You have to lift and separate. Sorry guys.

2. X-ray screeners fail to notice restricted items, let alone real threats. I know dozens of people who have accidentally carried scissors, nails, knives, and large volumes of liquid onto planes. Moreover, the TSA admits that it has difficulty detecting partial explosive devices. Screeners routinely fail bomb drills.

3. Metal detectors are still the principal line of defense at many airports. They fail to detect ceramic and glass weapons. You can pick up quite nasty ceramic blades at any cooking store. For that matter, metal detectors also can’t detect several types of weakly paramagnetic metals, like titanium.

4. Terahertz radiation has a 90% penetration depth of a few centimeters at best. That means it can’t detect objects concealed in body cavities, or, for that matter, implanted objects.

5. Even backscatter x-rays with a significant penetrating distance won’t be able to distinguish between medical implants and weapons–as the new software only reports anomalies for physical screening.

In addition, there are two systemic problems making threat prevention difficult:

1. Because passengers are not forced to undergo screening during transfers, the safety of the entire US air system is limited to that of the weakest connecting airport.

2. Because not all screening methods are applied to every passenger, and because the methods used can be deterministically altered by passenger action (e.g. refusing AIT screening, choosing to enter security during high load times, selecting which airport to enter), large threat classes that could be prevented can go unnoticed.

Even if all screening methods were used in concert at all airports, internal weakly-magnetic weapons would still go undetected. Yeah, it sounds crazy; but we’re talking about people willing to kill themselves on planes. Who does that? Even if we hadn’t prevented any of the publically reported terrorist attacks in the last ten years, you’d be more likely by an order of magnitude to die as a result of mechanical failure or pilot error than by terrorism!

I probably shouldn’t mention that queueing hundreds to thousands of people in small spaces before screening probably isn’t the safest way to do things, either.

Moreover, there are several types of legitimate objects which cannot be reasonably screened. Motorized wheelchairs, for example, can contain a hundred amp-hours at 12 volts, or 4.32 * 10⁶ joules. That’s about as much energy as a kilogram of TNT.

On top of all this, security personnel, pilots, and cargo aren’t held to anywhere near the same level of screening. The TSA, however, seems more concerned with suppressing criticism than actually preventing attacks. Our reaction to whistleblowers like Robert Cravens, who reported hundreds of pounds of flammable materials being stored improperly, or Chris Liu, who posted a video of airport security failures to youtube, has been, well, less than congratulatory. As their Human Resources department says, “because it is illegal to retaliate at the TSA there is no need to maintain an office for complaints.”

The funny thing is that these objections are obvious. I’m sure there are entire teams of people in the DHS and elsewhere fighting to improve the safety of flights, and that they are raising exactly these concerns. I can only conclude that their efforts have been buried by bureaucracy or dismissed due to conflicting ideologies.

I’m not trying to say we should do away with airport security. I’m saying that if we’re going to spend $8.1 billion annually, we might focus on more likely threats. We simply cannot prevent sufficiently determined attackers from killing people on planes. What we should do is focus on in-depth, comprehensive risk management.

That means taking a page from Israel’s book, where terrorist attacks actually happen on a regular basis. It means asking people questions, watching their behavior carefully, and other types of soft assessment. It means making it difficult to actually hijack the plane: reinforced cockpit doors, failsafe flight controls, and rapid scrambling of fighters. It means continuing to increase the presence of air marshals. We’ve made great progress in implementing these layers of security.

In fact, given that modern passenger aircraft are basically capable of taking off, flying, and landing completely under computer control, it seems entirely feasible (and indeed, I would be shocked if nobody were currently developing this) to make planes which simply cannot be hijacked or crashed into structures. At the first sign of hijacking, both pilot and copilot sign off on a failsafe landing mode. The plane does not permit manual override of its controls, finds the nearest airport, and touches down. A more passive layer of this software could simply prevent the aircraft from entering a flight corridor which could lead to collision with a major metropolitan area. Even if these safety modes are ten times more likely to cause crashes than manual control, it’s an improvement over the likely outcome of hijacking–and completely eliminates the hostage value of the passengers. There would simply be no incentive to commandeer a plane.

It also means means spending more money on good screeners, and training them to recognize more than the demo bombs on their screens. It means establishing stronger cultural and physical constraints on trusted employees, and raising the bar on background checks.

We can also mitigate threats through cultural systems. After 9/11, passengers are aware that they have significant impact on the outcome of a hijacking. We should encourage people to yell when their neighbors light their shoes on fire, and tackle them if they try to pull a weapon. It’s an experimentally proven and cheap way to prevent deaths.

In summary, the present TSA process fails to address a variety of realistic threats while placing undue focus on the specific attack modes we’ve already seen, at significant cost to the public. I disagree with this process, and opt out.