AI to Disrupt Psychology

Elvis Ruiz
8 min readJun 3, 2022

Social media is influencing people’s daily lives at an increasing rate. Meta, Instagram, Tiktok, and Twitter quickly became the most important stages for online social interactions. META (Facebook) is the most popular social network among the many Internet-based applications individuals use to communicate socially with others, with almost three billion users as of June 2022. META (Facebook), like other SMs' primary purpose, is to allow people to share their views and stories to form new relationships and maintain old ones. As a result, Facebook enables users to interact and communicate with people far more quickly than before. From this perspective, the link between META use and psychosocial outcomes has piqued the interest of researchers.

Social media has a positive and negative impact on how we see ourselves, and we must recognize the consequences to minimize their effect on our mental health and personality. One of the intriguing traits that might be evaluated for adaptation is personality. In the field of research, a person’s personality can be defined as a set of specifications that compel a propensity on the person’s conduct, which is constant over time and situations. Knowing of one’s personality period can provide insight into how he would react in various situations. Knowing a user’s personality can help you anticipate his wants, traits, and future behavior. As a result, adaptive applications may use personality models to read their behavior accordingly.


Photo sharing on various social networking sites (SNSs) has evolved into an essential aspect of the online social experience. SNS users exhibit their personalities, lives, and preferences through various forms of photographs, particularly selfies. Humans are social by nature; the online meta experience opens a window to a person’s personality traits and needs.

The psychological community has just recently begun to grapple with the reasons for sharing content on social media. According to an article titled “Why We Share: A Study on Motivations for Mobile Media Sharing,” respondents were asked to keep a diary of their posting behaviors and feelings and then participate in post-study interviews. After studying media sharing behavior, the researchers discovered “that social and emotional influences had a key impact in media sharing behavior” after studying media sharing behavior.

Some studies have looked into how social media has influenced children’s psychological development. According to a study published in the Journal of Experimental Social Psychology, people post on social media because it can lead to good social media feedback and self-esteem. More specifically, the desire for likes or followers on social media significantly impacts why people post. Many users are encouraged to share more socially due to the good attention they receive for posting.

In general, people publish from an emotional place where they are looking for a reaction. Because social media is based on communication, it’s only natural that the primary incentive for posting stems from a desire to connect with others. But this constant quest for acceptance and exposure on social media can lead to significant psychological problems for some.


No matter how idealized one depicts themselves on social media platforms, their social media accounts nearly always reflect their genuine nature. For many of us, the social networking site's selfies, faces, and layout might serve as personality forecasters. It’s crucial to remember that both online and offline activities are real. When doubt arises, one can rely on consistency to determine whether or not a behavioral pattern is genuine.

Because an individual’s personality can change depending on circumstances, qualities must be labeled. The Big-Five model is used to determine a person’s personality. Openness, Conscientiousness, Extraversion, Neuroticism, and Agreeableness are the five traits that make up the “OCEAN” model. The broad five-factor model is used in this study to measure personality qualities that divide people into five agents: openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism. Very extroverted individuals are warmer and more confident than calm and apprehensive. They publish vivid images compared to conscientiousness. Agreeable people are well-organized and pleasant. They share photographs with a color scheme that corresponds to extraversion. Individuals with high levels of conscientiousness are well-organized and exact. Rather than uploading images of items, they prefer to publish photos of people. Neurotic people are less likely to be emotionally resilient—people with neuroticism experience negative emotions. When posting grayscale photographs, they rank first out of all five attributes. People who have a high level of openness are receptive and prefer innovation to regularity. They upload photos that include sketches.

The OCEAN model can express as much variance in people’s personalities as possible using a small number of trait dimensions. In addition, qualitative traits which humans naturally perceive can be applied to ML models. These models can be used to build an AI capable of perceiving human personalities and even predicting future behavior. Simply put, a person’s personality may be determined by looking at the photographs that they publish on social media.

Everything counts. All data is relevant. Some pictures are lonely, some are enthusiastic, some profiles are darkly templated, some are mainly videos, and others are private, demonstrating that people are uninterested in displaying their photos and belongings in public.

A quick ML analysis using Tensorflow, Hadoop, Google Vision AI, and SM profiles shows that:

  1. Many public profiles do not primarily display images of a person but photographs and quotes that reflect their moods.
  2. Middle-aged folks seem to post on things that are on their minds, such as rapid technological development and comparisons between their and modern worlds.
  3. We can observe that people who upload nature photographs are interested in the natural world and enjoy talking about it.
  4. Profiles composed of mainly black and white photographs seem to correlate with some level of depression. Depressed persons are more likely to post images in blue, grey, and other dark colors. This can be linked and reinforced with sad images and negative quotes.
  5. A colorful WhatsApp profile and status could signify good mental health, while a melancholy soul could have a black and white image.
  6. Profiles composed of filtered photographs, which may make the actual image more vivid and brighter but do not match their vibe, maybe be a symptom of a gloomy individual.
  7. How you edit your selfies and how much you enhance them says a lot about your personality. Selfies that have been heavily manipulated imply that the individual may have low self-esteem or suffers from body dissatisfaction.
  8. Posting selfies after selfies may also be indicative of narcissistic and antisocial conduct. Narcissistic conduct can also be shown in frequent posts about personal accomplishments, nutritious diets, and physical fitness.

On their social media pages, people can be identified by what they post about social activities, intellectual themes, achievements, physical health, religious beliefs, and opinions on various topics. Surprisingly, one can use online behavior to estimate a user’s age. Most youngsters who erase their posts if they do not obtain enough likes may reflect a person overly concerned with self-approval. Teenagers also upload photographs based on their moods, whereas adult users prefer to write about various themes.


Artificial intelligence (AI), with machine learning, executes advanced, human-like operations and has the potential to revolutionize a variety of industries, including health care, transportation, education, finance, and more. At their finest, AI technologies can outperform people in terms of speed, scalability, and accuracy, freeing up time and resources for us to work on challenges that machines can’t. Combining approaches like data mining (creating new knowledge through deep study of massive amounts of data) and expert analysis, AI technology offers important tools for therapy. AI opens up the possibility of diagnosing current and potential issues and testing and confirming forecasts and solutions.

According to my research, artificial intelligence can extract better information about a person’s personality qualities from selfie images than human raters. Researchers from HSE University and the Open University for Humanities and Economics in Russia have shown that the technology may be used to locate the “best matches” in customer service, dating, and online tutoring.

Physiognomists have attempted to link facial appearance to personality since Ancient Greece, but most of their hypotheses have been invalidated by modern science. The few documented links between certain facial features and personality traits, such as the ratio of facial width to height, are minor. Human raters were asked to make personality judgments based on images, and the results were inconsistent, implying that our judgments were too inaccurate to be of any practical use. Nonetheless, there are compelling theoretical and evolutionary grounds to suggest that the human face can communicate some information about personality features, particularly those important for social communication. After all, face and behaviors are shaped by genes and hormones, and social experiences resulting from one’s appearance may affect one’s personality development.

However, current data from neuroscience reveals that the human brain interprets images of faces holistically rather than looking at specific facial traits. Researchers from two Moscow universities, HSE University and Open University for the Humanities and Economics, have collaborated with BestFitMe, a Russian-British company, to train a cascade of artificial neural networks to make trustworthy personality judgments based on photos of human faces. The model's results outperformed those found in earlier studies that used machine learning or human raters. Based on selfies the volunteers posted online, the artificial intelligence made above-chance assessments regarding conscientiousness, neuroticism, extraversion, agreeableness, and openness. The personality assessments that resulted were consistent across different images of the same people.

The research involved 12 thousand participants who completed a self-report questionnaire based on the “Big Five” personality model and uploaded a total of 31 thousand “selfies.” The participants were divided into two groups, one for training and the other for testing. A series of neural networks were employed to preprocess the photographs, removing faces with emotional expressions and pictures of celebrities and cats to maintain consistent quality and characteristics. Then, after training an image classification neural network to split each image into 128 invariant features, a multi-layer perceptron was utilized to predict personality traits using image invariants. The average effect size of r =.24 means that AI can correctly predict the relative standing of two randomly picked persons on a personality dimension in 58 percent of cases vs. the 50% expected by chance. In comparison to meta-analytic estimations of correlations between self-reported and observer assessments of personality qualities, this shows that an artificial neural network based on static facial photographs surpasses an average human rater who sees the target in person for the first time.

Conscientiousness was found to be more clearly recognizable than the other four personality qualities, according to the research ‘Assessing the Big Five personality traits using real-life static facial photographs,’ published in Scientific Reports. Female face personality predictions showed to be more reliable than male face personality predictions.

There are a plethora of potential applications to investigate. In circumstances when speed and low cost are more critical than accuracy, personality recognition from real-life images can supplement established techniques for personality assessment. Artificial intelligence can be used to recommend products that are a good fit for a consumer’s personality or to find the best matches for people in dyadic interactions like customer service, dating, or online tutoring.

In a later article, we will explore how ML is capable of providing a tangible grade to each important personality trait.