http://www.dailymail.co.uk/sciencetech/article-5197747/AI-detects-expressions-tell-people-lie-court.html

I think this is awesome. When the Brethren, middle management, missionaries, or anyone answers questions or gives talks (and even those pesky anti-Mormons) then we'll be able to better determine if they are telling the truth or not.

Now of course the most reliable way according to our glorious doctrines for detecting truth is via the Holy Ghost. But its sure awesome to have all these latter-day technological advances to help move the Truth forward more honestly. Thank Heavens that no leader/missionary would dare suggest that we be anything but completely honest/transparent or else people would begin to doubt we've got the True Church.

OK.
But once you know what it's looking for (and that's not hard, they tell us!), you can train yourself to not do those things when you lie.

Which would probably make it as easy to beat as a polygraph (which are essentially worthless).

:)

ificouldhietokolob Wrote:
-------------------------------------------------------
>
> Which would probably make it as easy to beat as a
> polygraph (which are essentially worthless).


Partner works HR in a major metropolitan police force, which successfully uses polygraph to staff sensitive positions.

Many fail and do not go forward; failures can be grieved but rarely if ever are.

But more than that, just knowing that part of successfully qualifying for a position requires a polygraph selects out those who couldn’t pass from applying in the first place.


But to your narrower point, that a tiny few individuals can successfully ‘game’ the test. True. But there is vastly more to the polygrapher’s skill and the polygraph exam itself than just the actual test. Most people don’t know what those things are and thus cannot anticipate them.

Polygraph testing while not perfect is far from being useless.

Are they actually "successfully" using it, or just think they are?

Because in reality, objective testing of polygraph "reliability" puts it at or around 50% -- in other words, tossing a coin to see if somebody is lying is *just as reliable* as a polygraph "test."

Which is why they're not acceptable as evidence in a court of law...because the science of testing them shows they're not reliable.

Where they might be considered to "work" is in scaring people...ignorant people who think they DO work will sometimes "fess up" (or avoid the test) if they're "guilty" of something, fearing failure. Some people consider that a "successful" use of a polygraph. I don't. It's "successful" use of fear and ignorance.

http://www.apa.org/research/action/polygraph.aspx



Edited 1 time(s). Last edit at 12/20/2017 02:15PM by ificouldhietokolob.

This reminds me of the story of the two coppers who wheeled a photocopy machine into the interview room where they were trying to get a guy to confess to a crime.

They explained the photocopier was a lie detector. They plugged it in and listened to it warm up. Unbeknownst (is that a great word word, or what!) to the criminal, they'd laid a sheet of paper on the platen with the word LIE! written on it in big letters.

They then asked him a question, he supplied an answer and one of the cops hit 'print', and out onto the feeder tray came a piece of paper with "LIE!" printed on it.

The criminal was stunned and (for the sake of brevity and because cops have short attention spans) immediately confessed.



Stories were a lot simpler back before cell phones and tech gadgets...

It's from an episode of The Wire. And it's hilarious. I highly recommend.

ificouldhietokolob Wrote:
-------------------------------------------------------

> Because in reality, objective testing of polygraph
> "reliability" puts it at or around 50%...

May I see some peer reviewed literature for this 50% claim? Also, bear in mind what I said, a polygrapher’s skill goes beyond the actual test.


> Where they might be considered to "work" is in
> scaring people...ignorant people who think they DO
> work will sometimes "fess up" (or avoid the test)
> if they're "guilty" of something, fearing failure.

From a police force HR perspective, this is rightly considered a successful use of polygraphing.

Anonbecauseobligated Wrote:
-------------------------------------------------------
> May I see some peer reviewed literature for this
> 50% claim? Also, bear in mind what I said, a
> polygrapher’s skill goes beyond the actual test.

From a study conducted for the Congressional Office of Technology Assessment:

"Neither available data nor theoretical analysis indicates that polygraph tests function as claimed by their proponents. Substantial numbers of both truthful and deceptive individuals
may be misidentified through use of polygraph tests,
and the tests can be "beaten." For most common uses of polygraph testing there is not even rudimentary evidence to support such use, and reliance on polygraph testing to protect national security would appear to be very problematic."

https://www.researchgate.net/profile/Leonard_Saxe/publication/232511382_The_Validity_of_Polygraph_Testing_Scientific_Analysis_and_Public_Controversy/links/0a85e53b9e29c9b7c9000000/The-Validity-of-Polygraph-Testing-Scientific-Analysis-and-Public-Controversy.pdf

That paper has references to studies giving "reliability" percentages, and the reasons to be wary of them -- that in fact, very few well-conducted studies have been done.

The "50%" number comes from David Lykken's scholarly book, "A Tremor in the Blood." You'll have to buy it.

> From a police force HR perspective, this is
> rightly considered a successful use of
> polygraphing.

That might be a successful strategy for getting people hiding things not to apply/finish their application, but it has absolutely nothing to do with polygraphy (as elderolddog's copy machine story shows nicely). Whether or not the machine does anything at all in that situation is irrelevant, as is any supposed "skill" of any examiner.

It also can't tell them how many unafraid, practiced liars didn't succumb to fear, and got on their force. Odds are very good they hired such people.



Edited 2 time(s). Last edit at 12/20/2017 02:56PM by ificouldhietokolob.

ificouldhietokolob Wrote:
-------------------------------------------------------
>
> That might be a successful strategy for getting
> people hiding things not to apply/finish their
> application, but it has absolutely nothing to do
> with polygraphy (as elderolddog's copy machine
> story shows nicely). Whether or not the machine
> does anything at all in that situation is
> irrelevant, as is any supposed "skill" of any
> examiner.

How do you mean it has nothing to do with polygraphy? My point is the same as elderolddog. The mere presence of the test creates more honest behaviour than otherwise.

Which goes to my initial post, which contested your claim that polygraphs are worthless.

Thank you for the link.

Human Wrote:
-------------------------------------------------------
> How do you mean it has nothing to do with
> polygraphy? My point is the same as elderolddog.
> The mere presence of the test creates more honest
> behaviour than otherwise.

One could say the same thing about a made up god.

"Don't do bad things or made-up-god will smite you."

See? The made up god is real!

Also Anon for this Topic Wrote:
-------------------------------------------------------

> One could say the same thing about a made up god.
>
> "Don't do bad things or made-up-god will smite
> you."
>
> See? The made up god is real!

I don’t know about “real”, but we sure do know it works, right?

No, the made up god doesn't work. The theater works. That means anybody who doesn't believe the theater is immune from the technique.

Instead of a polygraph, why not have a box with flashing lights and wiggling needles and bubbling fluid that attaches to a vibrating set of head-mounted electrodes. It would probably work better.

Do you know why polygraphers don't do that? Because they either believe in the pseudoscience and try to follow the traditional methods scrupulously, or they know it's BS and there's no point in changing anything.

Also Anon for this Topic Wrote:
-------------------------------------------------------
> Do you know why polygraphers don't do that?
> Because they either believe in the pseudoscience
> and try to follow the traditional methods
> scrupulously, or they know it's BS and there's no
> point in changing anything.


Either way, there’s still utility in polygraph testing, in certain contexts and with most people.


The larger point here is that the same is true with the OP’s AI system. Many will buy into it and so it will be effective with many people.

The crucial question is in what contexts are we willing to permit the use of them?

Anonbecauseobligated Wrote:
-------------------------------------------------------
> I don’t know about “real”, but we sure do
> know it works, right?

No, actually, we don't know that.
And in fact, evidence clearly shows it *doesn't* "work."
Since believers in such things fill up the jails at higher rates than non-believers in such things.

Oh, and it's way more than a bit of a stretch to give any "success" due to psychological manipulation to "polygraphy." Call it what it is, don't give the "success" to something that's irrelevant to the "success" or failure of the manipulation.

ificouldhietokolob Wrote:
-------------------------------------------------------

> No, actually, we don't know that.
> And in fact, evidence clearly shows it *doesn't*
> "work."
> Since believers in such things fill up the jails
> at higher rates than non-believers in such
> things.

Not sure what this means.


>
> Oh, and it's way more than a bit of a stretch to
> give any "success" due to psychological
> manipulation to "polygraphy." Call it what it is,
> don't give the "success" to something that's
> irrelevant to the "success" or failure of the
> manipulation.

A stretch, how so?

Most people believe polygraphy will out their deception. Therefore, most people will register the sorts of physiological responses polygraphy is designed to measure.

You seem to be arguing that since some can train themselves (or are born with little to no affect) to "beat" a polygraph, it is therefore an invalid measure for all people. Not true.


As for your point about "manipulation" etc, all that is also part of any polygraph test. They are part of the skill-set of the polygrapher, which you wish to dismiss.

It's not all about the machine, which is what you are focusing on. In fact, in some specific scenarios, the ones that concern you most (lack of affect, "gaming"), the machine to the polygrapher is not unlike a pack of Tarot cards to the fortune teller.

But in most cases, the machine is actually measuring valid data points. (I will read your link when time permits.)

What is of concern here with the OP, is in what contexts are we willing to employ AI polygraphy? I've been outlining an HR context, circumscribed by consent etc. I think these are valid and useful. But obviously the larger implication of the OP is that AI polygraphy can be used everywhere without any sort of consent etc.

There are two problems with this.

First, the fact that some people can be trained to beat a polygraph and others lack the physical responses on which polygraphy depends is actually a huge problem. For if you don't know who is immune to polygraphy before the tests, you can't tell if your results are valid or not.

Second, if polygraphy is in some senses the equivalent of Tarot cards, why not just use Tarot cards? That would at least be more honest.

I've always maintained that relying on pinkie swears holds the greater promise for the dawn of a new, truth-telling, day. And you can quote me!

I quote you frequently.

And sometimes they don't kick me out anyway.

> From a police force HR perspective, this is rightly
> considered a successful use of polygraphing.

What you're saying is, even though polygraphy is pseudoscience and the process itself is fraudulent (like EOD's photocopier story), the theatrical use of polygraphy in an interview has some utility.

That I can believe. But it doesn't mean that the polygraph itself works.

There is no method of lie detection that performs much better than chance. None. Not the polygraph, not voice stress analysis, not brain fingerprinting, not micro expressions, not fMRI. Nothing really works. These methods are championed by people who want to make money from them and their claims are repeated by people who don't know any better.

One interesting point to consider is that champions of one method will preach about how well their's works while decrying the other methods as garbage. It reminds me of certain churches who claim that they are the only ones who can save you, and that all other churches are false.

Also Anon for This Topic Wrote:
-------------------------------------------------------
> > From a police force HR perspective, this is
> rightly
> > considered a successful use of polygraphing.
>
> What you're saying is, even though polygraphy is
> pseudoscience and the process itself is fraudulent
> (like EOD's photocopier story), the theatrical use
> of polygraphy in an interview has some utility.
>
> That I can believe. But it doesn't mean that the
> polygraph itself works.

My claims clearly don’t go beyond that point.

As for claims about the the validity and soundness of the test itself, keeping in mind that the “test” is far more than the mere measuring of certain physiological responses with a machine, I won’t concede that it’s all just pseudoscience, but I will agree that there are flaws that should cause concern.

IS NOT ALLOWED IN THE COB, SLC, UTAH, OR THE IMW. period.

I have used polygraphs to make sure a client is telling the truth. What is interesting is that it helps because people will tell the truth when they think a lie would be caught. In some cases getting the whole truth makes it possible to convinces the prosecutor to reduce charges. It is important to be listening and watching because an incorrectly worded question may look like a lie in the answer. IOWs I always say that a bright polygrapher could make Mother Teresa as a liar. But with proper questions the answers are useful in finding truth. One must also evaluate truth by physical evidence. I have had a few lying clients but none to lie on the polygraph, though one was close.

There is a man in the WH who cannot distinguish truth from lies or fact from fiction. Indeed, he would be right more often if he merely flipped a coin.

I just retired from a career in which the assumption was that everyone was lying. Game on!

This presumption allowed a clever fraud to be perpetrated for awhile: lies against self-interest that were thusly accepted as truth, but which ultimately benefited the liar.

Man, do I have stories! (Boring stories, but stories nonetheless)

Ok I have deducted that it is not doctrinally possible that you worked for the Quorum of the Twelve. Every time we read the publicly transparent and totally honest and consistent statistical, financial, doctrinal, and historical records that they put forth in realtime we can be so touched with tears welling in my eyes over such awesome honesty and transparency. If it wasn't for this being a reality then people would suspect its all a fraud and that everyone should run away from it as fast as they can.

I'm skeptical.

I mean, the technology may work on most people--as polygraphs do--but there are systemic problems with polygraphs that will likely vitiate the new procedure as well.

One problem is that people can learn to pass polygraphs. All major intelligence services train their operatives on how to fool those exams. And the training often works.

Another, and in my opinion more important, problem is that sociopaths pass polygraphs all the time. They do not have normal fear/anxiety responses: that is part of the personality disorder. So they lie and their physical reactions very often don't rise to the level that the exams will catch.

So will the facial recognition technique work for the vast majority of people? Probably yes; and probably at the same reasonably high rate as polygraphs. But as with polygraphs, the people you most want to catch in deception are likely to escape detection at quite frequently. Anything that relies on physical reactions to stress will only work if the subjects have normal physical reactions.

I heard that same anecdote elder dog but you left out one detail.

The cunning cops placed a metal colander on the perp's head and affixed it with a wire to the copy machine.

Clear cut case of police brutality.



Edited 1 time(s). Last edit at 12/20/2017 02:53PM by Shummy.

the cops in your story must have been religious!!!

Isn't it true that all the wiles used either mechanically, electrically or visually rely on "tells"?

The notion is that when one tells the truth, the mind and body remain 'placid'. But when a lie passes the lips, the knowledge that it is not true creates the electrical, mechanical or visual 'cue' that registers as evidence of a lie.

So the natural or acquired ability to hide the cue is what we're talking about.

There are people who have honed their skills in spotting the visual and aural cues; you know who you are!



Edited 1 time(s). Last edit at 12/20/2017 04:28PM by elderolddog.