I can see how pressure would be applied when seeing the machine is leaning towards "lying", possibly breaking the subject's effort to lie.
But what would be the interviewer's strategy, if the subjects insist that they are telling the truth regardless of how the interviewer manipulates the machine? Wouldn't it immediately start discrediting the whole process if the subject is in fact telling the truth? I'm telling the truth here, and yet your machine says I'm not, hence it's broken, and hence I'll happily lie in the subsequent questions when it actually matters.
Most innocent people doubt their innocence when strongly accused even when they know they are right. Just a tiny bit, but in the right setting and with enough wearing you down, you can make innocent people believe they did it. I've seen it happen right in front of me.
But what's the point of making innocent confess to false crimes in this setting? (i.e. requiring polygraph for job application)
I would imagine the entire point of doing a test would be to find out who is innocent and who is lying about being innocent. If you pressure the innocent into false confession, wouldn't it just make everything even more difficult?
In the context of a job application? Making up excuses to hide discrimination, maybe artificially limit the pool of applications to get around other hiring restrictions? Put the applicant on the back foot or make them share information they normally wouldn't? There are probably many ways to abuse it.
The polygraph would be used as the “bad cop” in the good cop, bad cop routine. After a line of questioning where the interrogator/polygrapher suspects lying, or is fishing for more information, they might say something like “everything sounded good but the machine is showing some deception. Is there anything you can think of that might be causing these readings? Anything you didn’t tell me? I want to get you out of here, but we need to resolve these results.” If the machine does show spikes during certain lines of questioning but not others, for instance about someone’s timeline on the day of a murder vs their relationship to the victim, it can be a reason to pursue further questioning in that area.
Given all the ways polygraphs can be misused or abused, the only real use I see is as in interrogation tool. But given the issues with false confessions in general, I think the interrogation should hold less weight, but that is a whole other issue.
But what would be the interviewer's strategy, if the subjects insist that they are telling the truth regardless of how the interviewer manipulates the machine? Wouldn't it immediately start discrediting the whole process if the subject is in fact telling the truth? I'm telling the truth here, and yet your machine says I'm not, hence it's broken, and hence I'll happily lie in the subsequent questions when it actually matters.