Monday, May 31, 2004

Some in GOP want to axe TSA

Just a few days after my and photo and post documenting my many misadventures with the U.S. Transportation Security Agency (here), some GOP members of Congress want to privatize the TSA.

I'm not sure privatization is the answer. Privatization probably would improve performance, by allowing government to play contractors off against each other.

However, my understanding of 911 is that hijackers deliberately selected airports where screeners were unusually weak. It later turned out that many of these security firms were in parts of the country where it was apparently easier to comply with regulators by making campaign contributions to the "right" politicians than fixing their problems.

No doubt similar considerations will play into calls to privatize the TSA. The GOP would no doubt benefit from campaign contributions from security companies eager to get "government off the backs of business"; the GOP no doubt received such contributions prior to 911. Democrats, on the other, receive much of their money from unions; large government bureaucracies usually mean strong unions (as the gov't is not especially good at union busting) and a nice democratic constituency.

So the GOP wants private screeners while the Democrats want a large TSA bureaucracy, and it all has to do with campaign contributions rather than real performance.

Uniformity of training and auditing is probably required. A mix of TSA supervisors with private contractors do most of the real labor would probably be optimal, but TSA agents would still be required continuously at every security station in my view to prevent a repeat of 911 problems. That sounds like a larger TSA than the GOP envisions.

I think security screening is fundamentally flawed. We keep hearing that only 60% or so of simulated guns are actually confiscated by screeners. But there is only such one can do with X-rays or even manual inspection of baggage. Fundamentally, is a flawed process. If they only catch 60% of the guns, this still deters terrorism --- if they stop 3 out of 5 hijackers at the gate, that would sound alarm bells all over the airport (and all over the country), and every passenger on that plane would get a thorough going-over. That would break up the cell, and foil the plot, so terrorists won't risk it. But it is a flawed process, and 40% of the guns might get through, and that's probably enough to pull off a plot.

A particular problem in the U.S. prior to the use of secondary screenings, is that we effectively had a single secure area at all of our airports. Once someone got into the secure part of one airport with weak security, they can bring that weapon into the secure area of any other airport. So terrorists simply needed to find the airport with weakest security, get into the secure area by nook or crook, and then transport the weapon's pieces to the airport where they actually want to cause problems.

This is an issue at Pittsburgh, for example, where someone made the mistake of building a huge shopping mall inside the secure part of the airport. This was built prior to 911, when non-ticketed passengers were allowed into the secure part of the airport, so it was thought this was a good idea. The problem with letting non-ticketed passengers into the secure area, again, is that once security has been breached (by, for example, bringing small pieces of a weapon into the secure area using multiple non-ticketed passengers) it can be assembled and transferred into the secure area of any other airport. So, opening up the secure area at Pittsburgh brings the security at all other airports into question.

The only solution I see is to continue with random secondary screenings. However, these should be conducted in a professional manner, with plenty of time allowed for passengers to board their planes. I think the secondary screeners need to be provided with x-ray machines and metal detectors similar to the primary screeners, as the current practice has too many people up for secondary screenings.

Ultimately, for all my privacy concerns raised in my now famous "Pentagon Spy and Me" story, I do believe in data mining.

I believe the ultimate solution is similar to the one used by Israeli airlines.

I'm told Israeli airlines require all passengers to be pre-cleared by the government, usually involving an interview as well as a detailed questionnaire.

In the U.S., this would mean replacing the current CAPPS I and CAPPS II system, which frequently permanently tars people with names similar to know terrorists (including their Western-sounding aliases), with a biometric database system.

Frequent fliers and Visa applicants would be expected to interview with the government, which would establish an account with biometric identification. At the airport, they would simply to present their hand to a machine to get a reduced level of screening due to their pre-clearance. This would be coordinated with U.S. customs and the Visa application, so it would also accelerate entry and exit by foreign nations, eliminating the current fingerprinting system that many find so offense. Pre-cleared foreign nationals would interview once when applying for a Visa, and effectively fingerprint themselves (under the watchful eye of Customs) when entering and exiting --- much more user-friendly. Customs already had a similar biometric pre-clearance program in place, which it abandoned for reasons unrelated to security.

Would such a system be misused by the federal government, for example, to gain information about, or even harass, political opponents of the sitting administration? History teaches us that, without strong safeguards and independent supervision, it would absolutely be abused.

This is one reason why the Administration's current dabbling in (and misuse of) unsupervised data-mining programs (as I describe from person experience in Pentagon Spy) are so dangerous --- they can potentially completely discredit this technology, preventing a system like the one I've described that has the potential to make transit through both airports and customs faster and safer.