AI Outdoes Humans on Inferring Personality Traits From Facial Features
A new Russian study demonstrates that artificial intelligence (AI) is able to infer people’s personalities from “selfie” photographs better than human raters do.
The technology was able to make above-chance judgments on the “Big Five” personality traits — conscientiousness, neuroticism, extraversion, agreeableness, and openness — based on 31 thousand selfies the participants had uploaded online.
The personality trait of conscientiousness emerged as more easily recognizable than the other four traits. In addition, personality predictions based on female faces appeared to be more reliable than those for male faces.
The findings, published in the journal Scientific Reports, may have significant implications, as the technology can be used to find the “best matches” in customer service, dating or online tutoring.
Investigators from Ancient Greece to the Italian physician and criminologist Cesare Lombroso have tried to link facial appearance to personality, a practice known as physiognomy. But the majority of their ideas have failed to withstand the scrutiny of modern science.
The few established associations of specific facial features such as facial width-to-height ratio with personality traits are somewhat weak. Studies asking human raters to make personality judgments based on photographs have produced inconsistent results, suggesting that our judgments are too unreliable to be of any practical importance.
Still, there are strong theoretical and evolutionary arguments to suggest that some information about personality characteristics, particularly, those essential for social communication, might be seen in the human face.
After all, face and behavior are both shaped by genes and hormones, and social experiences resulting from one’s appearance may affect one’s personality development. However, the recent evidence from neuroscience suggests that instead of looking at specific facial features, the human brain processes images of faces in a holistic manner.
For the study, researchers from two Moscow universities, HSE University (Higher School of Economics) and Open University for the Humanities and Economics, have teamed up with the Russian-British business start-up BestFitMe to train a cascade of artificial neural networks to make reliable personality judgments based on photographs of human faces.
The performance of the resulting model was more accurate than those from previous studies which used machine learning or human raters. The artificial intelligence was able to make above-chance judgments about conscientiousness, neuroticism, extraversion, agreeableness, and openness. The resulting personality judgments were consistent across different photographs of the same individuals.
The research was conducted with a sample of 12 thousand volunteers who completed a self-report questionnaire measuring personality traits based on the Big Five model and uploaded a total of 31 thousand selfies.
The participants were randomly split into a training and a test group. A series of neural networks were used to preprocess the images to ensure consistent quality and characteristics, and exclude faces with emotional expressions, as well as pictures of celebrities and cats. Next, an image classification neural network was trained to break down each image into 128 features, followed by a multi-layer perceptron that used image invariants to predict personality traits.
The findings show that AI can make a correct guess about the relative standing of two randomly chosen individuals on a personality dimension in 58% of cases as opposed to the 50% expected by chance.
This indicates that an artificial neural network relying on static facial images outperforms an average human rater who meets the target in person without prior acquaintance.
Pedersen, T. (2020). AI Outdoes Humans on Inferring Personality Traits From Facial Features. Psych Central. Retrieved on August 3, 2020, from https://psychcentral.com/news/2020/05/25/ai-can-guess-some-personality-traits-from-facial-features/156806.html