Could some people pass a lie detector?

We put this to Jim O'Shea, Intelligent Systems Group, Manchester Metropolitan University:

Old-fashioned lie detectors like the polygraph only detect stress. If the lack of remorse meant that the interviewees had reduced stress levels that would help them pass. Our lie detector, Silent Talker, makes its judgement on non-verbal behaviour: crudely what people call body language. Silent Talker can detect stress but lying involves other factors. We can only juggle a certain number of mental variables at once while we're thinking. If we've got to try and maintain a whole load of different factors about an imaginary story it's very difficult to do all the mental processing to keep that consistent. That's what's known as having a high cognitive load which affects non-verbal behaviour. Also duping delight occurs when liars get a kick out of getting a lie across successfully and again this affects non-verbal behaviour. In one of our own experiments on the general population we taught silent talk to recognise guilty feelings the participants felt while they were lying. When we added this information to the general lie detection we got more accurate classifications. In another independent study conducted by a different university using Silent Talker it was found that Silent Talker was effective at detecting lies told by psychopaths in interviews. So there we have it: evidence that remorse is a factor in the general population but also evidence that in the case of one disorder it's not the only factor.