Can a machine detect if someone is unfaithful or has committed a crime?
A man undergoes a polygraph.
A man undergoes a polygraph.Dima Korotaev (Getty Images)

Machines capable of detecting if someone is unfaithful, commits a crime or lies in a court statement are part of the collective imagination thanks to multiple movies, programs and reality. While some celebrities like Cristina Ortiz, known as La Veneno, have faced polygraph tests on television, there are reality What love on surety either the game of your life that they use technologies to supposedly detect if contestants are lying. These types of devices, which in theory would have enormous potential in criminal investigations, have generated great controversy due to their error rates.

Everybody lies. Some do it about their relationships. Others to hide something, benefit or get money. Although many falsehoods are trivial, others can have devastating effects on society, employment, criminal justice, politics, public health and even national security, according to research published in the journal Applied Cognitive Psychology.

For humans, detecting lies is difficult. They usually do this with an accuracy of 54%, since they are often conditioned by a truth bias and tend to judge other people as truthful. This is indicated by the cited study, according to which police officers and other professionals in charge of making credibility judgments do not do much better. In this case, they experience a lying bias that leads them to think that others are lying.

Faced with the inability of humans to detect deception, multiple researchers have tried for decades to create technological solutions capable of doing so – although without much success. The first thing to keep in mind is that there are many different types of lies. Don Grubin, professor of forensic psychiatry at the University of Newcastle and expert in polygraph tests, emphasizes that a lie about an act (“I was not in that restaurant”) is not the same as one about a fact (“Paris is the capital de España”), an intention (I will go to the party) or another that is emitted by omission (not saying a critical piece of information).

“Depending on the type of lie, there are devices that aim to record brain changes or physiological responses associated with cognitive work and salience associated with deception,” he says. While some devices try to record active brain regions during lying or changes in the size of the pupil, others, such as polygraphs, evaluate the autonomic nervous system. The expert considers that the promises of many of the new technologies to detect deception are not justified: “They are based on very limited research and more on the marketing than in science.”

Do polygraphs lie?

There is also no scientific evidence that polygraphs can accurately detect lies, according to the American Psychological Association (APA). Some courts, including the United States Supreme Court, have repeatedly rejected the use of this type of evidence due to its unreliability. However, according to the APA, these types of technologies are sometimes used in non-judicial settings to assess workers, assess the veracity of reports from suspects and witnesses, and monitor some offenders on parole. For example, to people convicted of sexual crimes.

The instruments used to conduct polygraph tests typically assess three indicators: heart rate or blood pressure, respiration, and skin conductivity. To do this, a pneumograph (an instrument that records breathing movements) is placed around the chest of the person in question. Cardiovascular activity is monitored through a blood pressure cuff and skin conductivity is measured with electrodes placed on the fingertips.

These devices have generated quite a bit of controversy. Although several investigations indicate that they detect deception better than chance —with an accuracy that reaches 85% in some cases—, they still present significant error rates. “Far from being perfect, it’s a vast improvement on how people detect deception,” says Charles R. Honts, a professor of psychology at Boise State University and a polygraph examiner.

unreliability

The main limitation of these instruments is that there is no scientific evidence that there is a single physiological response pattern associated with lying. As the APA indicates, “someone honest may be nervous when he tells the truth and someone dishonest may not be anxious.” “The idea that we can detect the veracity of a person by monitoring psychophysiological changes is more a myth than a reality,” says the agency.

Added to this is the risk that examinees will try to manipulate the results of these tests. Some strategies used to pass polygraph exams can be effective and difficult to detect, according to research published in Journal of Applied Psychology. Among them, the authors mention biting the tongue, pressing the toes against the ground or counting backwards by sevens. In addition, according to the APA, some previous psychological interventions to manipulate the beliefs of individuals and the use of pharmacological agents that alter arousal patterns can also influence.

Although polygraphs have been touted by their proponents as a triumph of science, they also have their detractors. Psychologist Leonard Saxe, professor at Brandeis University and author of several investigations on polygraphs, assures that they are not valid means of detecting lies because “the scientific basis is poor”. If such tests were to be validated, erroneous results could have devastating consequences for individuals: some could be fired from their jobs, have their reputations damaged, or even end up in jail.

For some experts, it is difficult to imagine a scenario in which this type of technology is precise enough to be used in critical situations, such as trials for murder or terrorism. Despite this, multiple researchers do not cease their efforts to achieve it. While some seek the key to detecting lies in facial muscles, others explore eye movements, facial temperature, or the speaker’s words and gestures.

You can follow EL PAÍS TECNOLOGÍA at Facebook Y Twitter or sign up here to receive our weekly newsletter.


#machine #detect #unfaithful #committed #crime

Leave a Reply

Your email address will not be published.