With rapidly developing technology and increasingly imaginative criminals, there has been a rise in identity theft these days. As a result, one should be extra cautious and protect oneself from scams and other fraudulent scenarios by keeping their personal details secure. Because, if you become a victim of identity theft, your personal information such as name, maiden name, address, date of birth, social security number, passwords, phone number, e-mail, bank account and credit card numbers, could be used by thieves to exploit you.
What is worse that these details can be used by them over and over again. For instance, the security questions asked about personal details, like, “what’s your mother’s maiden name?” or “On which of the following streets have you lived?” can be easily interpreted by a driven identity thief.
So, how can one identify if it is a fake identity? While identifying fake profiles is a main problem in security, researchers in Italy have demonstrated a system that can detect whether the person answering the questions are genuine profile holders or not.The researchers have published the findings of their new study in PLoS One.
In order to carry out the experimental identity verification task, a group of 40 people from Italy were quizzed about their personal details by the researchers. The task involved the group to answer 6 expected questions, 6 unexpected questions, and 4 control questions that required a ‘YES’ answer. Similarly, the same amounts of questions were given to them that required a ‘NO’ answer.
While half of the participants were asked to respond to the questions honestly, the other half were asked to memorize details of fake profiles and use them in the quiz. The group that had to fake it was told to take a mock quiz twice before the test began to ensure that all the necessary information required to answer the questions are accurately memorized by them.
When the test began, the researchers used the MouseTracker software to keep a track of the movement of each respondent’s mouse while they answered the questions.
The quiz consisted of 12 questions like, “Do you live in Padua?” and “Are you Italian?” the answers to which an identity thief could easily remember and answer. However, the twist arrived when the respondents were asked to answer the unexpected questions.
For instance, in the second round of 12 questions, they were quizzed about “What is your zodiac sign,” which can be easily answered by truth tellers but can be complicated for the liars to determine.
“While truth-tellers easily verify questions involving the zodiac,” the study says, “liars do not have the zodiac immediately available, and they have to compute it for a correct verification. The uncertainty in responding to unexpected questions may lead to errors.”
The mouse-movement data collected from the quizzes was later analysed by a trained machine learning algorithm, according to which, the typical average movement of the mouse of the liars varied from the truth tellers when they moved the cursor from the bottom of the screen to the answers at the top. In fact, the researchers found that the overall dishonesty of the liars was infecting their movements even they were telling the truth and they could accurately be identified as lying. The researchers found out that 95% of the time, the system was able to distinguish the fake responses from the real ones.
To test if the model can proficiently classify participants from different cultures, the entire experiment was tested on 20 German subjects (10 liars and 10 truth-tellers). However, the end outcome was the same here too. They write:
“From a cognitive point of view, what is interesting here is that, in the experimental design, the mind-set of the liars also extended its effects to questions when they were responding truthfully. To our knowledge, this pattern of results has never been reported before and could be an indication of the level of sensitivity of the technique of mouse-movement analysis.”
It is too early to comment if such a system can be used as a reliable method for identifying identity theft, as the sample tested by the researchers were limited in size. However, this new system has opened way for more and more researches in this area, which coupled with fine tuning of algorithm, could one day make this system a trustworthy method.