Scientists can now predict personality and success level with a selfie

Researchers at the University of Pennsylvania have devised a way to scan people's faces to estimate personality traits and real-world success, and, according to the scientists behind the method, it actually works.

2

A new study has found that machine learning techniques can identify correlations between facial characteristics and real-world success, leading to the creation of predictors based on facial characteristics for specific aspects of life. The team used an AI system trained to estimate the likelihood of the Big Five personality traits from human face images.

The researchers took headshots from 96,000 MBA college graduates and fed them into the AI. The AI produced results indicating the likelihood of each student's personality traits: neuroticism, conscientiousness, agreeableness, extraversion, and openness. The team then reviewed the AI's results and evaluated how each student's career had panned out, and found a correlation between the AI's predictions and each student's success.

• Read more: Experts left terrified after AI self-replicates without human assistance

The real-world application of such technology could be devastating, not to mention unethical, as companies could simply adopt an AI system to screen job applications. Essentially, applicants would be judged purely on how they look by an AI system designed to estimate a person's success based on appearance. Taking that example even further, banks could adopt it to decide whether to grant a loan or even issue health insurance.

"Widespread adoption of facial recognition technology in the future may motivate individuals to modify their facial images using software or even alter their actual appearance through cosmetic procedures," writes the researchers behind the study