It's really confusing, you might be wondering right now about all that anxiety you had before entering a job interview, you've grown up with!

Or so you think.

You are now sitting in a room with white walls, everything else in it is also white except for the black leather chairs that sit on one of them, the place is large but claustrophobic, you almost swear that for a moment you have smelled the smell of medicines or disinfectants because you feel that you are in a hospital, and while you are drowning in Your daydream Someone calls your name, it's your turn.

Well, the decisive moment has come and the dream job has approached you, it remains only to show some confidence and speak with a degree of fluency and tell them that you would like to sit in their place after five years, but the surprise was that they began to tell you that the personality detection algorithms that examined your posts, photos, and some videos You posted it on your Facebook account saying that your face gives sharp hostile signals, and it also says that you are a person who tends to dominate others and is less open than required, which makes you not suitable for this job.

You might think that what we talked about a while ago is a bitter idea for one of the interesting episodes of the famous series "Black Mirror", or perhaps as one of the scenes in a science fiction story that over-imagines the future of the world, but the capabilities of artificial intelligence to reveal characters through data and images The face has evolved over the past few years in a way that raises serious questions about the future of human beings with their presence, simply asking: Can this really happen?

To answer this question, let's start from the year 2015, when (1) "Wu Youyou", a researcher at the Cambridge Center for Psychometrics, and her team, were able to develop a software mechanism that could describe people's personalities based on their "likes" for Facebook pages and posts. More than 86,000 people underwent experiments through a smartphone application called "My Personality" that evaluates their personalities and is allowed access to their Facebook data. After that, the application user is asked to invite a friend, coworker, or family member to evaluate his personality. .

Here came the results (2) to say that this mechanism was more accurate in describing the "personality" of individuals from their colleagues at work, their friends, and family members, and the more data the program obtained, the more accurate the results. But what is remarkable is that this smart mechanism needed very few "likes" to characterize people. For example, in the study, the computer was able to predict the personality of an individual more accurately than a coworker by analyzing only ten likes, more accurately than a friend with 70 likes, and with more accuracy than a family member (father, mother, and siblings) with 150 likes, What is more striking is that the computer was able to describe the features of a person's character more accurately than his wife, only by examining 300 of his Facebook likes!

The personality tests that the participants underwent in the experiments were based on defining the "five (3) elements of personality," which are openness (whether a person is curious, open to the world, or introverted into himself), and conscience (which is the degree of a person's ability to control himself in the face of commands. Instructions and the performance of duties and important matters), social sense (the person's ability to interact with others and trust in himself during this interaction), compatibility (trust in others and thus treat them well and cooperate with them accordingly), and disorder (degrees of anxiety, negativity, and depression of the psyche in general).

This is not the first time that the Cambridge Center for Psychometrics has succeeded in using user data for Facebook and Twitter in analyzing their personalities.In 2013 researchers (4) from the same center succeeded in developing a new algorithm that identifies personality, race, gender, IQ, sexual orientation and political orientation. For people, whether they are addicted or not. In these experiments, artificial intelligence succeeded by more than 80% and up to 95%, in a sample of 58,000 people, not only that, but the artificial intelligence was also able to determine the religion of the subject subject to the tests with the same degree of accuracy, Muslim or Christian.

Can this be believed? At first glance, it calls for laughter, artificial intelligence, just by analyzing a relatively small number of your likes on Facebook posts and pages, was able to find out your self-confidence rate, your openness to new experiences, your level of social mixing, the amount of sadness that inhabits yourself, and how you adjust By yourself, more than your wife! But the signs of laughter in our faces as a result of this paradox may disappear when we meet Michael Kosinski from Stanford University, one of the participants in that study with "Wu", but Kosinski, in the year 2017, made a completely new leap in this field.

Kosinski believes in only one thing, which is that everything is possible for artificial intelligence, which has the ability to maintain and access huge amounts of data, as well as the ability to analyze it using algorithms more complex than the capabilities of the smartest human beings, when you decide - as a human - to analyze an issue, A problem that you face in your life, you rely on only one example and two, and often we humans tend to be biased and inclined to irrational mechanisms during our attempts to analyze things. As for artificial intelligence, it is more capable of avoiding bias, to the extent that it can see things that you cannot see. Things like sexual preference!

(5) The artificial intelligence mechanism developed by Kosinski and his team, after examining photos of 35,000 people on a dating site, in which people mention the sexual orientation of the partner wanted for dating (gay or heterosexual), has a 91% success in determining sexual orientation. Just by checking five pictures of the same person. For initial experiments, this is an amazing result after which we can easily expect close to 100% in an opportunity not so far away.

Although the idea of ​​Kusinski seems simple at first glance, it depends mainly on the position that people take during filming (especially for the purpose of dating) and the way clothes and hairstyles are styled and how to shave the chin, which causes gay men to adopt feminine postures and facial expressions, and vice versa for lesbians. The basic idea is the ability to process huge amounts of data, a natural person cannot easily read these signs, as for artificial intelligence, it has a greater ability, in large stages, to find patterns, through deep learning, then compare them with each other, and arrive at results with this accuracy.

Kosinski pushes the matter further, especially when (6) this research team announced their ability to develop artificial intelligence with deeper algorithms that can, just via your image on social media without any other data, determine some other criteria such as political orientation. And IQs. For us, the human face is a window to contemplate their feelings while we talk together, but for these researchers the human face is an open book haunted by a huge amount of data that only needs someone to come up with a simple self-learning algorithm and read it.

To understand this idea of ​​an "open book", we can leave Stanford for a while and go to the German city of Stuttgart, specifically the "Machine Learning and Robotics Laboratory", where Sabrina Hope worked with a joint international team to develop a new algorithm that tries to identify only people's personalities. Through their eye movements, in their own experiments, subjects wore special glasses that can track eye movement, and then each of them was sent on a simple task to the store to buy some products and then return to the headquarters of the experiments.

That period was enough for the cameras to record a short video of the movement of the eyes, and then the algorithm that works through deep learning determines the five basic features of the subjects of the experiments at a rate that exceeds the barrier of natural coincidence by a difference of 15%. To understand this idea, we can consider a recent study from the University of Melbourne (8) in Australia in which researchers were able to develop a biometric camera that, only when you stand for seconds in front of it, can determine 14 traits for you, ranging from age and gender to the five basic personality traits.

It seems that we are on the verge of a completely new world in which some people can know "everything" about you, just from "likes" or a picture or a video.

Could this happen? That you pass a regular camera at the entrance to a company and the security guards refuse you entry because the artificial intelligence mechanism suggests that you are a very aggressive personality or that you will be rejected in a job without the need to meet you, but only because analyzing your picture on Facebook gives signals that you are not suitable for the job? We do not know, but we know without a doubt that developments in this area of ​​identifying people's personalities through their data or - what is more profound and important - their images on social media are developing at an incredible speed, especially as it is strongly supported by the Ultras (9, 10, 11). A huge economist wants to chase everything related to the customer, besides the desires of the intelligence services that reach the point of sensuality to know the conditions of people, this prompts us to expect anything to happen in the future.

The Egyptian popular proverb says "You know so-and-so? Ah, you associate with him? No, he stays with you," indicating that knowing people's personalities and their essence does not appear to us until after a long period of interaction between us and them, but it seems that we are on the verge of a completely new world that some can To know in it "everything" about you only from "likes" or a picture or a video. Of necessity, this leads us to a very important question: Who can these technologies fall into? At first glance, we envision that there will be useful uses for this accelerated development towards that type of technology, but on the other hand there are wide possibilities as well. These technologies can help tyranny, control people's desires, classify them as classes only by tracking the features of their faces.

In the famous movie "Enemy of the State", Robert Dean, an unknown lawyer, tries to escape with a murderous secret that he knew by chance related to an influential man in the US administration, but he was surprised that the US intelligence services were following him in a way that he could not believe. Something watched: the phone, the television, the doors and the clothes, the satellites follow him wherever he goes, the cameras in the shops, the metro, the traffic lights and everything else, could one day come the state, or perhaps a company or a coalition of companies, as close as possible to Zeus, in the ancient myth King of gods who watches the finest minutes of the world from the mountains of the Olympics and knows everything that is going on in them?

____________________________________________________________

Sources:

  • Computers using digital footprints are better judges of personality than friends and family

  • Computer-based personality judgments are more accurate than those made by humans

  • The 'Big 5' Aspects of Personality

  • Digital records could expose intimate details and personality traits of millions

  • Deep neural networks are more accurate than humans at detecting sexual orientation from facial images.

  • Face-reading AI will be able to detect your politics and IQ, professor says

  • Eye Movements During Everyday Behavior Predict Personality Traits

  • Biometric Mirror highlights flaws in artificial intelligence

  • Massive AI Twitter probe draws heat map of entrepreneurial personality

  • Psychological targeting and mass persuasion

  • Neural Network Learns to Identify Criminals by Their Faces