Egads; the brainscanner is on the horizon. This is scary stuff.
The MRI-based devices purport to measure the lie itself. Sections of the brain light up on the scanner the moment a lie is formed in the mind. It sounds like science-fiction, but the machine is already working its way into the justice system.
Don’t even think about lying. The “brainbox” in question is officially known as a functional magnetic resonance imaging (fMRI) detector, a medical device which companies in the US are now marketing as a lie detector, claiming accuracy of 90% to 100%.
Now anybody that has 1/1024th of a statistically-minded brain is going to read that and freak out. Assuming that the marketing claims aren’t completely off base (and really, when are all marketing claims based entirely upon reality), this means the device has an error rate as high as 10%. Let’s pull out some quick numbers, shall we? A quick google search shows one reference that says that as of June 25th, there were 581.2 million passengers in 2007. Assuming that’s accurate, that means its not unreasonable to say that there are in the ballpark of 1 billion (say it with me, do your best Dr. Evil impersonation, “beeeeelion!”) instances in a year of people taking trips on airplanes. Do you see where I’m going yet?
Let’s say that the device is five nines accurate. So it’s not 90% accurate, it’s 99.999% accurate. For every 100,000 individuals, we’re going to have one false positive. Well, you say, that doesn’t sound so bad… until you remember that we have 10,000 x 100,000 passengers in a year. Well, now, 10,000 people held up because they didn’t pass a screening is a giant number. Some people may shake their heads and say, “Well, that’s the price we need to pay to be safe.”
Remember, it still only happens one out of every 100,000 passengers. So for individual TSA screeners, it’s going to be a fairly uncommon occurrence – an individual screener is going to process how many people in a year? Someone failing one of these tests is going to be guilty in the eyes of the screeners. They’re not just going to be held off of a flight, they’re going to be interrogated. “The device is 99.999% accurate, they must be guilty!” They’ll be put on “no-fly” lists. If they’re off their medication, they might be tasered. If they’re packing an iPod, they might be shot.
Now imagine that it’s actually only 99% accurate. Now for every 100,000 passengers, we have 1,000 false positives. Out of a billion travelers, we’ll have 10,000,000 false positives. TEN MILLION false positives. You can’t even interrogate that many people! At least here you won’t have the problem of over-reacting… if one out of every 100 people trips the brain scanner, the TSA screeners will stop paying attention to the damn thing altogether (think how many people are in line during the holiday season; it would take days to get through security).
Now, some people will say, “Well, we won’t scan *everybody*, we’ll only scan *certain* people.” So now instead of holding up 1% of travelers, we’ll just be holding up 1% of people the TSA thinks ought to be scanned. When you’re talking about the number of people involved in air travel, trying to reduce your sample isn’t going to help much.