Innocent Until Proved Guilty? (Voice Stress Lie Detector Exposed as Sham)

ABC News ^ | March 30, 2006 | Brian Ross

[George Maschke]

'Innocent Until Proved Guilty?' Questions Are Being Raised About a Voice Lie Detector Used by 1,500 U.S. Police Departments and the U.S. Military

March 30, 2006 — - A Pentagon study obtained by ABC News finds that a new kind of voice lie detector used by the U.S. military and American police departments is no better than "flipping a coin" in detecting lies. Until the Pentagon ordered a halt to its use, the Voice Stress Analyzer was being used by military intelligence interrogators at Guantanamo Bay and in Iraq. Several suspected terrorists were released from custody based on the machine's results and former Iraqi deputy prime minister Tariz Aziz was one of the many "high value targets" who were hooked up to the now discredited machine.

Watch "Primetime" tonight at 10 p.m. ET for a full report from Brian Ross.

A laptop, microphone and software program make up what is called the computer-voice stress analyzer, or CVSA. Used by police departments across the country, this lie detector is a foolproof system to help catch criminals and liars, according to the man behind it.

"Police departments have paid $10,000 per system over the last 18 years and rely on it exclusively for truth verification," said Charles Humble, chairman and CEO of the National Institute for Truth Verification, which sells the CVSA. "We have a remarkable record of success."

But as questions surrounding the scientific validity of the machine and Humble's credentials grow, not everyone agrees that that should be the case .

[George Maschke]

Click on the link to read the full article. There is also a webcast.

ABC News Primetime exposes the Computer Voice Stress Analyzer, manufactured and marketed by "Dr." Charles Humble (who got his degree from an unaccredited diploma mill) and the so-called "National Institute of Truth Verification," as a fraud. It is a national embarrassment that ntil recently, military intelligence personnel were using this quack device to divine whether prisoners were telling the truth.

For related reading, see the CVSA Discussion Forum on AntiPolygraph.org.

[The Red Zone]

The chief use of such pseudoscientifery is to con confessions.

[George Maschke]

Unfortunately, many voice stress test operators actually believe the device works.

[Recovering Ex-hippie]

And what if I talked like Donald Duck? Hmmmmm.

[Air Force Brat]

Uh huh.

"Until the Pentagon ordered a halt to its use, the Voice Stress Analyzer was being used by military intelligence interrogators at Guantanamo Bay and in Iraq. Several suspected terrorists were released from custody based on the machine's results and former Iraqi deputy prime minister Tariz Aziz was one of the many "high value targets" who were hooked up to the now discredited machine."

No mention of the possibility for innocent Americans to be railroaded by this pseudoscience. I've never trusted lie detectors. From what I've read, they're too easy to game and/or misinterpret.

[Thud]

All these doo-hickeys are very culture-specific. The interrogators, interrogees and sensor interpreters have to be of the same culture and speak the same language.

[George Maschke]

But even then, they simply don't work. Lie detectors, whether voice stress analyzers or polygraphs, are a fraud.

[babygene]

My company, wrote and marketed a voice stress analysis program in the early 90's. We seriously developed the product based on research that was available at that time from the government.

The product we sold was "Inspector Gepettio Knows". It was marketed nationally, and although we gave it a cute name and had a picture of an adult Pinocchio on the box, we (anyway) considered it a serious product. And we made some money from it.

It was shipped (on floppy disk) with wave files of Nixon saying he's not a crook, Clinton saying he never inhaled, and Ollie North (I forget what he said). You would submit these "calibration files" to the program to demonstrate that it did indeed work.

Interestingly, a lot of copies were purchased by the Fed's and by the Israeli's.

It was written to run under 16 bit windows and has never been updated to 32 bits as this would have required us to buy a 32 bit Fast Fourier Transform library. The big problem we had in those days was the millions of calculations we needed to do on a 166mhz computer.

It actually runs faster today than then, even though it runs in 16 bit mode.

Freepmail me if you want a copy.

[Strategerist]

Basically, my understanding is lie dectectors only work if the subject THINKS they work.

Easily beaten. Not sure why professional organizations have so much faith in them.

The new frontier seems to be brain-scan lie detection, which actually is credible.

[babygene]

"But even then, they simply don't work."

You seem to have a pretty ingrained opinion on this.

Of course they work. Their just not infallible. Just as you can sometimes tell when someone is lying to you; by their voice, their face, their mannerisms... Sometimes it's clear, sometimes it's not.

[Hank Rearden]

I've worked on classified projects, traveled all over the world to sites with classified areas, worked in classified crypto environments - and the only polygraph test I've ever had to take in my entire life was when I was 16 years old. It was a routine screening by McDonalds, checking to see if we were really throwing away expired-time cheeseburgers, or eating them on the sly for lunch.

Ahhhh, priorities.

[Larry Lucido]

I still like the metal collander over the head, with wires going to a copy machine, with the machine spitting out a page that says "Lying!" in bold letters after each answer.

[George Maschke]

You seem to have a pretty ingrained opinion on this.

Not ingrained. Informed. There is absolutely no peer-reviewed scientific research indicating that these devices reliably work at better-than-chance levels of accuracy, or that they improve on the results that could be obtained by an interrogator without the device.

[babygene]

"There is absolutely no peer-reviewed scientific research indicating that these devices reliably work at better-than-chance levels of accuracy, or that they improve on the results that could be obtained by an interrogator without the device."

Your statement here seems contradictory. Is it no better than chance, or no better than a human can do? An interrogator can obviously so better than chance... YOU could do better than chance.

[George Maschke]

What I am saying is that there is no peer-reviewed research supporting either the notion that 1) CVSA "testing" works any better than chance or 2) interrogators using CVSA achieve better results than interrogators not using it.

[DH]

I still think that the old Indian method of "the finger in a bowl of water" truth detector worked as well, if not better, than anything made today.

Simply place the person's finger in a bowl of water and watch it make ripples as they lie.

[Erasmus]

If I recall correctly, it was an episode of Homicide: Life on the Street in which the detectives sat the perp down next to a Xerox machine. One of the detectives had two sheets of paper behind his back, and he could surrpetitiously slide either one onto the glass. In large letters, one said "TRUTH" and the other said "LIE." After asking the perp a question, the other detective would simply push the "analyze" (i.e., "copy") button. A few seconds later, he'd pull the "printout" off the machine and it would say "LIE!" The perp realized they had his number. He broke down completely.

The point of all these gizmos can't be their actual accuracy; it must be the psychological effect of this supposedly hi-tech stuff on the subject.

Unfortunately, I fear I'm thinking too rationally. The cops probably believe in this stuff just as much as the perps.

< ]8^(

[George Maschke]

If I recall correctly, it was an episode of Homicide: Life on the Street in which the detectives sat the perp down next to a Xerox machine....

I didn't see the episode, but it's mentioned in the snopes.com article, "Next Case on the Court Colander."

What's ironic is that when it comes to detecting deception, "real" lie detectors are little better than the legendary colander attached to a copy machine. Far too many in our law enforcement, military, and intelligence communities still believe in the myth of the lie detector. But America's jihadist adversaries aren't fooled. See, for example, the jihadi article (translated from the Arabic), "The Myth of the Lie Detector."

[Mikey]

Surprise: Police find drunk people in bars -- and arrest them.
by Bill Winter

Here's something to keep in mind next time you visit a bar, club, or restaurant: Police in several states have launched crackdowns on drunk driving and public intoxication -- and are now arresting drunk people who aren't driving and who aren't in public.

In at least two states -- Texas and Virginia -- police have started going into bars to arrest people who fail sobriety tests.

Police say the action is necessary to prevent drunk driving. They also say they don't have to wait until people leave a bar to arrest them for public intoxication, since the legal definition of "public space" includes the inside of drinking establishments.

The move has sparked outrage from civil libertarians, who say police are grossly exceeding their authority, and are arresting people who pose no danger to anyone.

This issue was first publicized in January, when WorldNetDaily.com reported that police in Virginia were going into bars and taverns and literally "pulling people off barstools." The police, dressed in "SWAT-like attire," would give people sobriety tests and arrest them for public drunkenness if they failed the test.

Then, on March 15, NBC Channel 5 TV in Dallas/Ft. Worth reported that Texas Alcoholic Beverage Commission agents and Irving police had launched a similar crackdown.

Police targeted 36 establishments and sent undercover police officers to look for people who appeared to be drunk. After administering sobriety tests in the bars and clubs, they charged 30 people with public intoxication.

In both Texas and Virginia, police said the campaign would reduce drunk driving. But police apparently made no effort to check which patrons had walked to the bar, or rode with friends, or planned to take a cab. In Texas, police even arrested people in a hotel bar who were registered hotel guests and had no plans to drive anywhere.

One Texas Alcoholic Beverage Commission agent said police were justified in arresting people because "going to a bar is not an opportunity to go get drunk." Another spokesperson said drinking can make people do stupid things like "jump off of balconies trying to reach a swimming pool and miss." The spokesperson did not give details about how prevalent the missed-swimming pool problem is in Texas.

After people complained, Texas legislators said they would review the program to "check for abuses" and to measure its effectiveness.

Well, here's one reason to oppose this campaign against social drinkers: While police are harassing tipsy people in bars, real criminals are walking the streets -- free to kill, rob, and rape again.

According to 2004 FBI crime figures, less than half of violent crimes in the U.S. are "cleared" by police. The FBI says only 46.3% of murders, forcible rapes, robberies, and aggravated assaults end with a suspect being arrested and charged. To break it down by crime, only 62.6% of murders, 55.6% of aggravated assaults, 41.8% of forcible rapes, and 26.2% of robberies are ever solved. This means that for the average violent criminal, there's better than a 50-50 chance he will never be caught. But hoist one extra beer in a bar in Texas or Virginia, and a SWAT team will nab you for sure.

If police spent fewer hours in bars and more hours investigating murders, rapes, and robberies, there's a good chance that more dangerous criminals would spend more time behind bars. And doesn't investigating violent crime seem like a better use of a policeman's time than lurking in bars, arresting your next-door neighbor for having one drink too many?

Sources:

(Special thanks to Hugh Wright)

http://worldnetdaily.com/news/article.asp?ARTICLE_ID=30288
http://www.nbc5i.com/news/8034788/detail.html
http://worldnetdaily.com/news/article.asp?ARTICLE_ID=49397
http://www.fbi.gov/ucr/cius_04/offenses_cleared/index.html
http://kutv.com/topstories/local_story_081200713.htmls

I guess next one won't be able to enjoy a cold one in one's yard after mowing the lawn. Get ' cha for mowing while intoxicated.

[Political Junkie Too]

I bought one of those Sharper Image "Truth Machine" voice stress analyzers as a toy. I keep it by the television and turn it on whenever politicians are on. It's fun to what the LED go red when they are asked tough questions that they are trying to dodge. It also makes watching C-SPAN more interesting.

-PJ