Lies about Lie Detectors
(May 20th, 2009) The growing market for "scientifically valid" means of detecting when you're telling a lie has become lucrative. But how valid is the science behind the detection of spoken lies? Two Swedish scientists have denounced the pseudo-science behind the gadgetry and are currently facing a legal battle for their honesty, reports Jeremy Garwood.
Commercially available "lie detector" machines are now routinely used to quiz potentially fraudulent claims for social aid and insurance payments, security screening and criminal investigations. Delegating such important decisions to a machine implies a sound scientific basis for this technology. However, when Anders Eriksson and Francisco Lacerda investigated claims made for lie detection technology, they found the field to be so bad they entitled their article: "Charlatanry in forensic speech science: a problem to be taken seriously." (
The International Journal of Speech, Language and the Law, 2007, 14 p. 169-93).
Unfortunately, one of the named "charlatans", Amir Liberman, an Israeli entrepreneur, has threatened legal action over the paper's findings. He has succeeded in scaring the journal's British publishers, Equinox, into withdrawing the article from the journal's website. In the latest twist, the Swedish Research Council, who funded the research by Eriksson at Gothenburg University, and Lacerda at the University of Stockholm, has sent a statement of protest at this "infringement on academic freedom".
So how valid are their charges of charlatanry? Or, more to the point, just how far has this discredited technology succeeded in infiltrating security-obssessed societies, ever more suspicious of their citizens' activities?
The Swedes set out their case: "A lie detector which can reveal lie and deception in some automatic and perfectly reliable way is an old idea we have often met with in science fiction books and comic strips." However, "it should be stated right away that at the present time no method for reliable lie detection is known and it is not even known if it should be possible to develop such methods in the future." Despite this, there are several products on the market claiming to work as lie detectors, albeit with other names like "stress or emotion analyzers".
They investigated the two most widely used types of lie detectors, the so-called voice stress analyzers (also referred to as psychological stress evaluators) for which they found "no demonstrable validity", and a newer type of analyzer said to be based on a multiple layer analysis of the voice. The latter comes in many different shapes as commercial products but are all based on "layered voice analysis" and the Swedes' troubles came when they reported that the validity of this technology "is to be found at the astrology end of the validity spectrum"!
These observations resulted in legal threats from Nemesysco, the Israeli company that sells the machines based on "Layered Voice Analysis" (LVA).
Nemesysco makes impressive claims for its LVA, saying that it has "unique signal-processing algorithms that extract more than 120 emotional parameters from each voice segment" classifiable into 9 major categories of basic emotions with "8 final analysis formulas". But Eriksson and Lareda noted that far from being widely published and discussed findings, such "extraordinary discoveries" in the field of human emotion "are completely unknown to the scientific community".
Contrary to the claims of sophistication, LVA turns out to be a fairly simple (and not very well-written) computer algorithm that analyses the relative frequency of "thorns" and "plateaus" that appear on the graphical display of a digitised speech signal. The program code is part of the 2003 patent documents and may be downloaded from patents on-line. Written in Visual Basic by Liberman, the company's director, the entire program comprises no more than 500 lines of code. "With respect to its alleged mathematical sophistication, there is really nothing in the program that requires any mathematical insights beyond very basic secondary school mathematics. To be sure, recursive filters and neural networks are also based on elementary mathematical operations but the crucial difference is that these operations are used in theoretically coherent systems, in contrast to the seemingly ad hoc implementation of LVA."
However, in his patent Liberman makes great assertions for LVA's statistical powers - "a crLIE value to 50 may be deemed indicative of untruthfulness, values between 50 and 60 of sarcasm or humor, between 60 and 130 indicative of truthfulness, between 130 and 170 of inaccuracy or exaggeration, and values exceeding 170 indicative of untruthfulness".
"And that is all there is," retort Ericksson and Lareda, "there is nothing special with these computations, except that there is no theoretical basis for them or independent motivation for the proposed ranges. The creation of digitization artefacts is completely independent of what the recorded sound represents. The program would analyze any sound the same way, be it a man speaking, an idling car engine, a dog barking, or a tram passing by."
Since the correspondence between the labels and what they represent "is perfectly arbitrary", they note that exactly the same LVA has also turned up in "special applications like the so called 'love detector'." Indeed, "why waste time and energy inventing a new program when all that is needed to build a 'love detector' is to rename 'cognitive stress' as 'sexual attraction'?"
In a rather unusual section, entitled 'Who is Mr Liberman?', the researchers describe their collaboration with a Swedish journalist who in 2004 actually tracked down Amir Liberman to a small office in the Israeli town of Natania. When asked about his academic background, Liberman admitted he had no university degree, but that he had taken some courses in marketing at an open university.
"To sum up by saying that there is absolutely no scientific basis for the claims made by the LVA proponents is an understatement. The ideas on which the products are based are simply complete nonsense."
But Liberman's LVA system sells well. In the US, it costs $25,000 for a laptop computer installed with his 'amateurishly' written software and the training necessary to use the system. Overall the market for these "lie detectors" runs to tens of millions of dollars. "What still remains for us to understand is how insurance companies, security agencies, and police departments can be willing to invest hundreds of thousands of dollars, pounds, and euros in equipment without ever asking who are behind the products, what are their qualifications, what are the scientific principles upon which the products are based. For us this is the real puzzle," lament Eriksson and Lareda.
Because lie detectors have been used in police investigations, by insurance companies and banks, and their use is increasing. For example, the UK Work and Pensions Secretary, John Hutton, apparently unaware of the severe doubts about this "technology", announced in 2007 that "lie detectors will be used to help root out benefit cheats. Voice-risk analysis software will be used by council staff to help identify suspect claims. It can detect minute changes in a caller's voice which gives clues as to when they may be lying."
And ironically, as Eriksson and Lareda explain, there is in fact one way in which the mere concept of lie detectors may actually appear to bring some success: thanks to the 'Bogus Pipeline Effect'! This is when subjects may answer more honestly if they believe that the truth can be tested for accuracy. Several studies have shown that just telling people that a lie detector is being used during their interview will make them more honest. "Whether the lie detector actually does anything or is even physically present is irrelevant."
Sadly, this is not only "morally dubious business". Thanks to the results of this charlatanry, people are being accused of crimes and fraud, facing criminal investigations, being denied insurance money and social payments. Even worse, there are plans to use such equipment in areas such as airport security where "since the LVA technique is totally unreliable it would mean that part of airport security will be based on decisions no more valid than throwing a pair of dice!"
