Because professional spies are trained to defeat the device; because pathological liars do not cause its needles to spike; and because our counterspies relax when a potential suspect "passes" — the system breeds the opposite of security.
Here's how I learned about that. In 1981 there was a brouhaha about the Reagan campaign having pilfered a briefing book used by Jimmy Carter to prepare for a debate. James Baker, to deflect suspicion from himself, hinted that it must have been the doing of the campaign chairman, Bill Casey.
Casey, just appointed C.I.A. chief, told me he was going to challenge Baker to a polygraph test to show who was lying. Figuring my old pal Casey was the culprit, I wondered why he would take the gamble. He reminded me he was an old O.S.S. spymaster, and that by using dodges like a sphincter-muscle trick and a Valium pill, he could defeat any polygraph operator. Baker wisely did not take Casey up on the challenge.
A more serious example of the foolishness of dependence on the machine: A national security adviser was suspected of leaking a secret to The New York Times. Though not our source, he flunked the exam, and was about to be fired and disgraced. He put President Reagan on the phone to The Times's publisher, who — on a one-time basis — confirmed that the adviser had not been our source. That was one fewer career lost to the predatory polygraph.
So Safire admits that his "old pal" Casey was a liar and had stolen Carter's briefing book, something which helped Reagan "win" a debate with Carter and contributing to his election victory (although not as much, probably, as Casey's (and possibly George Bush the First's) efforts to have Iran hold the hostages until after the election, known as the "October Surprise"). Safire then reveals how the NY Times sort of revealed a source, in the negative, to save the job of one of Reagan's national security advisers. Amazingly, Safire considers this to be a more serious case than his old pal's successful efforts to steal the presidency.