Hacker News Reader

Comment by puppycodes | Hacker News Reader

The only way polygraphs work is by convincing people its an actual lie detection machine. Cops leveredge this belief and tell you that you "failed miserably" so you ultimately confess because "your caught".

They are about as accurate as flipping a coin.

View on HN

Sure, but that's an argument against them ever being used in the screening process they use for people joining, not to only exempt specific high-level appointees.

That's beside the paint of political appointees getting special treatment. Patel's FBI has been using polygraphs extensively to test political loyalty

https://www.nytimes.com/2025/07/10/us/politics/fbi-polygraph...

> according to two people with knowledge of the questions and others familiar with similar accounts

After all the examples of the media getting caught playing fast and loose with anonymous sources, I’m going to take this one with a huge grain of salt.

DHS and Musk’s people confirmed it. See Reuters back in May 21.

From Teuters: "officials quickly launched an investigation to find out who was leaking to the press"

Leading is a bit different than "loyalty tests" isn't it?

The story is about “new applicants” and positions with no access to classified material. It is not illegal to share unclassified material, which is why DHS was having NDAs passed around.

> DHS said it was "unapologetic" about its drive to root out leakers. "We are agnostic about your standing, tenure, political appointment, or status as a career civil servant - we will track down leakers and prosecute them to the fullest extent of the law,” a DHS spokesperson told Reuters.

Maybe that’s true? I think HN was ok with this because it felt like a rational reaction to bringing guys like Big Balls into the SCIF.

It’s not illegal for me to share non-confidential data of my employer either, but if you were hiring, would you want to understand an applicants attitudes towards such leaks especially after having issues with it in the past?

I would.

So it’s not involved in a criminal investigation. Then it’s a loyalty test. Specifically, it reassures them you can keep a secret.

They are being asked if they have disparaged Patel or the administration. Not just keeping secrets. They want loyalty to the regime, not the nation. That is obvious enough given the overt actions of the entire DOJ.

To be fair, that's an important qualification for the job regardless of who is president.

It's just the political loyalty that's a bad requirement.

Why would political loyalty be bad? These are cabinet positions, not representatives of the voters whose actions change with changing voter sentiment.

The cabinet is chosen to implement the President’s policies. It would be odd to select for people opposed to the policies or those would actively try and stop them through leaks.

Political loyalty is not bad (to a degree) but it is never good when it replaces actual competence as the main qualification, or when the required degree of loyalty is unreasonable.

> They are about as accurate as flipping a coin.

They are actually worse than a coin toss.

The TV news show, "60 Minutes" tested several companies that provided polygraph testing services.

The show claimed some one stole some equipment. For each testing service the show said they suspected a different employee. In every case, the polygraph operator claimed they detected deception in the person they had been told by the show was the person already under suspicion. Polygraph is total bullshit, just used to add a pseudo-scientific shine to prejudice.

That CBS 60 Minutes exposé aired 40 years ago, yet it remains relevant today. It may be viewed in its entirety here:

https://antipolygraph.org/blog/2007/01/30/cbs-60-minutes-exp...

Which is a distraction from the fact that taking one is part of the process, and for no reason senior appointees have been exempted.

It would be a different matter if the entire process was being removed as policy, but nope, that's not it.

I don’t think either one is a distraction. Corruption at the highest levels of government is a serious problem. Government reliance on pseudoscience is a serious problem.

This is not accurate.

Polygraphs are not detecting lies, they're used to assess your sensitivities; there are really talented interrogators in counterintelligence, whose full-time job is to fuck with you in subtle ways. To poly a person at will is very much a power move, and some guys fucking love it. But that's a different story all together. Most of the time it's a formality like everything else. In reality, people don't have remotely enough bandwidth to pursue stuff like that unless there's a genuine investigation. But office politics people will office-politique.

Unpopular opinion: private companies should poly people more often in hiring, it could prove more useful than other arbitrary kind of culture fit interviews. Food for thought.

> private companies should poly people more often in hiring, it could prove more useful than other arbitrary kind of culture fit interviews.

Useful in what sense? That you can't figure out anyway what tested person is capable of because tested person can believe that they have skills on godlike senior level, but they are junior at best?

Yes I’m also curious. From what I know polygraphs and similar interrogation are for assessing whether there is anything that could be used to blackmail or compromise you. Whether one agrees with the method, the goal seems logical for intelligence orgs. For companies, industrial espionage would be the obvious parallel. I don’t know how polygraphs would relate to culture fit though… watching to see if candidates perspire and their heart beats faster when asked if they have grit and value diversity :)

Well, my perspective is having to do with Ukrainian mil tech, but I believe the same is true in e.g. cybersecurity, and it helps that the newcomers are often familiar with the process. Poly is not for everyone. Now, when you say "industrial espionage" it sounds dandy and fine, but for all intents and purposes there is no such thing. There is exploitation, yes, and it may easily cross from industrial, to defence, to political depending on trade.

I never said I subscribe to the unpopular opinion, but it's a point of contention regardless!

Thankfully, it is illegal for private companies to do that.

Unless you work in a pharmacy. Or you’re a ‘mall-cop’. Or literally any employee anywhere who is suspected of fraud or embezzlement or any “incident that resulted in a specific economic loss to the employer”.

You are correct. The Employee Polygraph Protection Act of 1988, which otherwise prohibits the use of putative lie detectors by employees, provides exemptions for such cases.

Unfortunately, that doesn't really prevent companies from doing things being illegal if they turn out to be profitable enough. You could use a multispectral hidden camera and an mmwave radar fed into 'AI' to simulate a lie detector - you can definitely get pulse and breathing rate out of it, probably also perspiration..

Sure. And then someone who set that up will get fired and leak the scheme, and nudes from mmWave will be found, and it’s all lawyers and liability.

> Unpopular opinion: private companies should poly people more often in hiring, it could prove more useful than other arbitrary kind of culture fit interviews. Food for thought.

Because what we need is more inscrutable judgement calls in the interview process?

Unpopular because it’s a bad idea. You’re now hiring better liars and scaring away humble talent.

And if you stop doing it, you now have a group who were polygraphed and a group who were not.

> Unpopular opinion: private companies should poly people more often in hiring, it could prove more useful than other arbitrary kind of culture fit interviews. Food for thought.

OK. I'll bite. I'm interviewing with you to be a frontend developer. You have me hooked up to a polygraph. What do you ask?

“Have you ever used a <table> for non-tabular data to avoid looking up the CSS grid syntax?”

“If you had the opportunity to inject pig-butchering scenery into our banking app, would you flee the country after its executed?”

> "This is not accurate"

I don't get what you thought was inaccurate?

I couldn't disagree with you more about your unpopular opinion.

I've been hiring for years and imo the best interview is a trial period doing the actual work with the actual team, not cop movie cosplay.

> whose full-time job is to fuck with you in subtle ways.

And this is a process that you expect to produce an output with any predictive value what so ever?

> it could prove more useful than other arbitrary kind of culture fit interviews.

You could also just end up selecting for psychopaths and sociopaths for whom this test does not function, regardless of how much you "fuck with them."

I did paint a cynical picture, so please do take it with a grain of salt.

Poly is just one tool, and it's never a solution to anything. What you say about potential abuses, undesirable selection, etc. I agree 100% and hate the office-politique jockeys as much as the next man. In most corporate environments, fucking with people has no place and use, but it's far from only application, in fact it's pretty much the worst application possible. Doesn't change the fact that it happens, even though it shouldn't. I also believe that it's important for many orgs to start taking security seriously; in the world where there's so much exploitation, and where applicable, it helps a lot if your org can have some kind of counterintel function. It doesn't have to include poly, but whatever helps you better understand the people you work with is surely a boon.

There is something of value it produces:

The relationship between politics and finance has lost accountability. Would you take a chance investing with just any cryptocurrency? Or will you make the wise decision and choose mine, the first and only crypto that’s polygraphed all programmers for accountability.

That's not as accurate as flipping a coin. It's a certain success or a probabilistic failure. A random coin flip could get you the wrong answer either way. But here, if someone confesses to something, you can be sure it's true. (unless you pass the threshold where people would rather lie that they did bad things, but that's a different problem... if you're getting the same results as torture, the interviewer is the problem, not the fake machine)

Reply on news.ycombinator.com