Detecting faked identities is a major issue in security. Usually, a large number of migrants from the Middle East entering Europe lacks documents, and their identity information is often based on self-declaration. A number of terrorists providing false identities is believed to be hidden among them. For example, the Berlin attacker of December 2016 was travelling from Tunisia and entered Germany and Italy from Greece using several false identities. Another resounding case was that of one of the suicide bombers that struck outside the Stade de France in Saint-Denis during the November 2015 Paris attacks. A passport of a 25-year-old man coming from Syria was found near the terrorist’s body. Investigators reconstructed that he entered Greece from Syria using a faked identity. One of the Brussels terrorists travelled through Europe using the identity of a former Inter Milan football player. Currently, as in the cases reported above, when the individual’s DNA or fingerprints are not included in a database there are very few techniques to spot a false identity.
Behavioural measures of deception are a possible tools for dealing with this investigative issue, but current lie detection techniques require the true identity to be known in advance, and therefore cannot be applied when such identity is unknown, as in most investigative cases.
The first lie detector in history was developed by psychologist Vittorio Benussi in 1912, at the University of Padova. Now a group of researchers from the same institution just reported, in the scientific journal PLOS ONE, a new technique aimed at identifying a false identity when the true one is unknown. Giuseppe Sartori, a lie-detection scientist and Professor of Forensic Neuropsychology at University of Padua, developed together with Merylin Monaro and Luciano Gamberini a novel technique that requires the suspect to undergo a computerized test in which questions about the claimed identity are presented on a computer screen. The subject is required to respond using the mouse: its trajectory is first recorded in its spatial-temporal characteristics, and then classified by machine learning methods as coming from a truthful or deceptive responder.
The researchers found that the analysis of the mouse trajectories, when the subject is engaged in responding to questions about identity, may be used to uncover whether the response was a truthful one or a deceptive one.
Giuseppe Sartori reports:
We have shown that kinematic analysis of mouse movements is a reliable index of the mental processes that underlie the production of a faked response to questions about identity. When a subject is lying about his identity, his trajectory is less direct and more erratic in directing the mouse pointer to the response button. The respondent intention to lie produces an atypical mouse movement. We have shown that this trajectory may be used to identify a liar with an accuracy over 90%.
The researchers have shown that when a subject is lying about his identity, the trajectory, velocity and direction of the mouse movement changes. While the truth-teller’s mouse goes directly to the appropriate response button, the liar’s trajectory is firstly attracted towards the button of the response that he wants to hide and then directed to the response button that corresponds to the faked identity.
The new methodology can be used with any computer, even for remote testing, and could be used as an identity-screening test for those cases in which objective evidence about identity is lacking. It could be a new, simple, and cost effective tool for security agencies, since it can run on any standard desktop computer with a mouse.
The original reasearch “The detection of faked identity using unexpected questions and mouse dynamics” appeared on Plos One.