Wednesday, October 12, 2005

A More Sensitive Approach to Terrorism

A new screening test can detect vanishingly small trace amounts of substances like cocaine and explosives.

Bad news for terrorists and drug traffickers: The hunt for narcotics, explosives and biohazards is about to get faster and easier thanks to new research from Purdue University.

A new testing method can, for the first time, speedily check objects and people for traces of chemical compounds....[T]he research from Purdue...developed a technique called desorption electrospray ionization, or DESI, that eliminates a part of the mass spectrometry process, and thus speeds up the detection of substances to less than 10 seconds....
There are some fascinating possibilities here. As Common Sense Technology notes, we may all be drug traffickers:
There is one study indicating that up to 97% of all bills in circulation in the country are contaminated by cocaine, with an average of 7.3 micrograms of cocaine per bill.
That would seem to be well within DESI's range, since it's successfully recognized chemicals at the picogram level.

The opportunities such a system could conceivably present for practical jokes, acts of revenge or malice, or planned disruption of air travel are also worth considering. But the main issue is false and accidental positives. Purdue's press release dances around this issue:
[T]he team's forthcoming spectrometry gear, which will weigh less than 25 pounds, fits into a backpack and returns a negligible number of false readings, both factors that are also important to law enforcement officials.
Apropos of false readings, Common Sense Technology makes a good point about random drug screening by dogs that have a 98% accuracy rate:
[I]f a dog and handler team maintains an accuracy rate of 98%, "[t]his means that whenever drugs are present, the dog will alert 98% of the time" and "whenever drugs are absent, the dog will not alert 98% of the time." Assuming that 0.5% of the population has drugs in their possession, if the "dog sniffs 10,000 people, 50 will possess drugs. Out of these 50, the dog will correctly alert to 49. Of the remaining 9950 people that do not possess drugs, the dog will falsely alert to 2% of this group, resulting in 199 false detections. Out of this population of 10,000, the dog has positively alerted to 248 people, 49 of which are correct detections and 199 are false alerts."
It'd be interesting to apply similar calculations to the DESI system, given that the sample population is much, much larger, and the terrorist population much, much smaller.

Unfortunately, I can't do that, since I have no definition of "negligible." Regardless, I'm keen to play around with some numbers. If the system delivered a false positive for explosives 0.001 percent of the time, that'd result in roughly 290 false positives a year at JFK Airport alone. I'd imagine that farmers, who are more likely to be contaminated with trace amounts of, say, ammonium nitrate or organophosphates, might come in for particular scrutiny - as might people who are taking sublingual nitroglycerin for a heart condition - but this is sheer speculation on my part.

It does seem safe to say that if the system also screened for drugs, the number of false positives would be higher...possibly a lot higher. Intentional efforts to overwhelm or confuse the system are another concern; it sounds as though one could bring an airport almost to a standstill with an ounce or two of the right material.

We'll see, I guess.


Samurai Sam said...

I'll believe it as soon as I see it on CSI: Miami.

Anonymous said...

RE: Cocaine traces on cash, does this mean I have to wash all my bills before I go to the airport? ;^D
G in INdiana