Criminals are becoming increasingly creative as technology advances. To avoid the loss of data and funds through fraud or scam, you need to ensure that personal details are safe. Identity thieves are likely to use your social security number, date of birth, email address, passwords, full name or maiden name, bank account, credit card, phone numbers, and other personal details to access your locked platforms, exploiting your sensitive data or money. It has become something to dread since the general public started using the internet. Many are concerned about this happening to them, yet we become nonchalant about the security questions we get asked to set up when creating new accounts, and log-in details.
A good case study is the 2015 Internal Revenue Service (IRS) hacking in the USA. Researchers found that hackers took their time to gather personal information about thousands of Americans, which helped them tackle security questions to gain access to their tax returns.
The hackers passed through all security protocols because they had provided the correct answers. However, if the system had put measures in place to detect ‘how’ the hackers gave the correct answers, there is a chance they would have realized there were lies flying all over the place.
An identity thief recognizes the process of unearthing your personal details has multiple layers. They will be patient, and once they get information about you, they will try to take advantage of it to get more. This might take a long time, even years.
Sometimes, Google asks you to key in some squiggly word, known as a ‘captcha’ (or Completely Automated Public Turing test to tell Computers and Humans Apart), for the system to confirm you are not a robot. You might have been wondering how this works, but it is all about the manner in which your cursor moves. Apparently, robots and human beings move the mouse with very different patterns.
The scariest thing is that identity thieves can use the information they have about you again and again, for instance, in using simple security questionssuch as ‘What is your surname?’, ‘What is your pet’s name?’ or ‘Which is your favorite car?’ That is why you need to go for more complicated security questions. Is it possible to discover an identity theft? Yes, but the process is rigorous, time-consuming, and tiresome.
Analysis conducted on mouse movement does not indicate each person has a unique pattern of mouse movement, but that the patterns can help in detecting lies.
Data security technicians have been working hard to develop more effective and reliable ways of detecting and preventing data theft on the internet. Over time, they have come up with interesting ways of doing this, thanks to Artificial Intelligence (AI).
Italian researchers have developed a system that recognizes whether any activity on your account is from the true profile owner or a fake. The personality identification experiment involved 40 Italians, with researchers asking them questions about their personal details. There were six unexpected and six expected questions, with more control questions requiring a ‘YES’ for an answer.
Half of the participants were to give answers about their true accounts, while the remaining half had to memorize details of various accounts belonging to other people. Those answering the quiz with fake accounts had to go through two mock tests before the actual one, ensuring they met all the required standards.
Once the test began, the researchers used software for tracking mouse movements. The software recorded each participant’s mouse movements. The questions were quite simple, things that identity thieves could easily remember, such as ‘are you Italian?’
A surprising twist of trends occurred when the participants had to answer unexpected questions. For instance, it was not easy for those representing fake accounts to say what their zodiac signs were, whereas someone telling the truth would not hesitate to give a confident answer. Some of the fake replies may need computing to offer an answer, in an atmosphere of uncertainty that leads to a high possibility of errors.
Analysis of mouse movement data via machine algorithms revealed mouse movement variations between truth-tellers and liars, especially from the taskbar towards the answers positioned at the screen’s top. One interesting thing is that mouse movements affected liars to the extent of showing they were not telling the truth, even when they were being honest. The system successfully set aside truth from lies, with a 95% consistency.
The researchers went ahead to test if the system could distinguish different cultures, with ten lying and ten honest Germans, reaching similar conclusions. Additionally, the fact that a liar’s mindset affects their mouse movements gave them a new challenge; exploring more cognitive correlations between mouse movements and cognitive authenticity. According to them, these trends are uncharted routes, because no one has ever noticed them before.
Later, questions about the practicality of the system arose, considering that the study involved a limited sample, meant to represent a proportionally large number of people, if the project was to succeed. Despite this and a few other challenges, the researchers thought the project had been a major step in the right direction, especially because it opened up concepts of algorithm fine-tuning, projected to help develop a method that can be trusted in the future.