You look for suspicious responses, but people are different from one another. Maybe a suspect is mentally disturbed, or merely weird? Maybe his response will be different from that of other people of his nationality, because his parents were immigrants?

“In such a case, we ‘ll see a very different response from that of other people of his nationality, but we won ‘t see an exceptional response to the special words. His responses to all the words will be equally odd.”

Can a terrorist develop a response to beat the system?

“Terrorists don’t know our special words, which we constantly change, nor do they know the ‘reasonable response ‘, against which they ‘re compared. It will therefore be very difficult for them to train. A terrorist can try to teach himself to mute his responses, or even take a tranquilizer. But in that case, his response will be different from that of other people of his nationality.”

What if a terrorist doesn’t chose his native tongue, in order to avoid the special words?

“He will respond differently from the way the average English-speaker responds, both because it isn’t his native tongue, and because he won ‘t understand some of the words. It ’s impossible to outsmart this machine; anyone who tries to be clever is screwed.”

What about someone who’s very interested in terrorism, but isn’t a terrorist? Maybe he ’s a member of Pakistani security?

“It’s possible, although not at all certain, that such a person would be a suspect within the 4% permissible range. If he is, he ‘ll have to explain himself to the human interrogator.”

How is the system tested? After all, you can ‘t collect terrorists to consent to being tested.

“So far, we ‘ve only tested and calibrated the system on Israelis. As I said, the goal is to create a system that tests the fear of being caught, so some of the subjects were given a mission, and others weren’t. The system was able to identify the subjects who were given a mission. We later specifically tested stressful missions, such as taking something from the boss ’s office. The system identified the people carrying out the stressful mission.”

The Department of Homeland Security allows you to suspect 4% of legitimate people. What is your margin of error in the other direction. Is the system certain of catching every terrorist, or will one periodically get through?

“We believe that maybe one out of dozens of terrorists can slip through the system. It’s not perfect, but that’s how it is with security – you can only do your best. You want things to make as difficult as possible for terrorists. If 50 terrorists are caught and one gets away, we’ll be pleased, and so will the authorities.”

The field is very hot, and there must be other companies mulling similar ventures.

“As far as I know, there are no systems that even come close to working on the same principle. There are systems that try to spot a lie through voice frequencies, but it hasn’t yet been proved that voice is an effective physiological index of intent. Another system is trying to spot excitement through body temperature, but what if the terrorist has taken a tranquilizer? None of these systems is as specific as ours. Even if someone were to read this article and decide to imitate us, he would face at least a two to three-year development hurdle.”

What’s your next step?

“Luckily, we’re in an all-or-nothing field. The Americans know that if they order the system only for JFK Airport, all the terrorists will switch to another one. The closure must be either hermetic, or nonexistent.

Let us hear your thoughts: Suspect Detection Systems Inc. Message Board