There’s a new contender for the Holy Grail object: The Magic Machine that Can Tell Truth from Lies.
On the face of it, this one seems even more useless than the old-style polygraph. It can be beaten by the simple expedient of “answeringquicklywithouthesitation.”.
The Times reported that psychologist Aiden Gregg has developed:
A new lie detector test shows that it takes on average 30% longer to tell a fib than to be honest.
That sounds an impressive test for truth – objective quantifiable, replicable, easy to measure, and so on.
Gregg said he built the test because he suspected that criminals were finding increasing ways to hide their dishonesty. …..
… The psychologist warned that existing lie detectors such as polygraphs – which monitor physiological changes such as blood pressure and body temperature – implicate too many innocent people. (from the Times)
Government funding for security is so reliable in these cash-strapped times for universities. So, in one way, it’s a great idea, from an academic’s perspective.
But I can’t see anything in this report that backs up its claims as a Holy Grail Machine.
The experiments were done in an environment which was not pressured. Completely unlike a real-world instance, subjects would have no reasons to be anxious about telling either lies or truth. However, thinking up experimental “lies” would mean subjects had to take more time than the took to tell non-lies.
If you were an innocent suspect sitting in front of one of these machines, for real, you would be worried about your answers. You might hesitate before saying anything, as you pondered possible implications. On the other hand, if you were guilty but had practised a good story, you could just reel it out. Quickly.
This machine might work for finding out which of a group of scared twelve-year-olds had graffittied the bus stop. (Although, elementary normal investigation skills would surely achieve that more time-effectively and actually produce valid evidence.)
Practised liars are convincing. They can smile and wail and even sob convincingly, witness Karen Matthews’ performances. The time-delay counting machine would never have uncovered what was true or false in what she said. Any innocent mother, in the position that Karen Matthews pretended to be, would not answer normally. She would fail the test, while the sort of person who could lie about such an event to their closest family and friends would probably come across as being truthful.
Flawed as this whole lie-detector machine concept is, you can pretty well guarantee that politicians will NOT welcome it unless they are confident that they can beat it easily.
So, if it does get the government go-ahead after its trials, you can at least be confident that it doesn’t work at all.