Computers Have Learned To Make A Psychological Portrait Of A Person By Eye Movement - Research, News

“Eyes are the windows of the soul,” we have always been told. Now it turns out that they are also a tool for revealing your character: computers have learned to scan eye movements and make a psychological portrait of a person

The joint experiment of the University of South Australia, the University of Stuttgart, as well as the German institutes Flinders and Max Planck Informatics was based on self-learning machine algorithms that correlate certain movements of the pupil with certain personality traits. For example, eye movements can tell artificial intelligence how social, consistent, or curious a person is. Using the derived algorithm, the software reliably recognizes four of the five main personality traits: neuroticism, extraversion, benevolence, and conscientiousness.

Scientists studied in detail the lives of 42 participants in the experiment "through their eyes" in the literal sense of the word. Everyone was engaged in everyday activities on student campuses, communicated with loved ones, solved everyday tasks while the sensors read the necessary parameters. Then everyone was asked to fill out questionnaires and identify the type of character. And after that, computer intelligence found stable relationships.

Dr. Tobias Letscher, one of the authors of the study, assures that the revealed connection between eyes and personality traits can be very useful in various aspects of social life, where robotics will be able to determine the signals given by a person in the blink of an eye (again - literally!).

“The value of the study is primarily in the fact that we studied the participants in their familiar environment, that is, they showed themselves as naturally as possible, which could not have happened in a laboratory. As a result, thanks to self-learning software, we got an algorithm that will allow the computer to recognize what kind of person is in front of him, predict needs, and often even predict mental disorders.

In addition, this algorithm can be useful in developing a more human and personalized approach to customers in industries that already use computerized communication schemes (chatbots, etc.) Programs will be able to recognize social signals and respond more flexibly to customer requests. In fact, this is a real breakthrough, a bridge in communication between man and machine."