Artificial intelligence facial recognition can predict political orientation from expressionless faces, according to a new study.
Researchers from the Stanford Graduate School of Business asked 591 participants to complete a questionnaire about their political beliefs.
Portraits of the group were shown to human raters who attempted to assess where they fell on the political spectrum, based solely on the shape of their face. They were also scanned by an AI algorithm.
The results showed that the accuracy achieved by humans and the algorithm was comparable.
The team says: "It was relatively high. In fact, it was on par with how well job interviews predict job success, or alcohol drives aggressiveness. The algorithm’s predictive accuracy was even higher when it leveraged information on participants’ age, gender, and ethnicity."
To avoid bias in the study, published in the American Psychologist journal, researchers made sure participants removed all jewelry, make-up, and shaved facial hair. Their hair was pulled back using hair ties, hair pins, and a headband while taking care to avoid flyaway hairs.
The study says psychological traits can shape facial appearance, with factors including facial care, diet, substance use, physical health, injuries, exposure to sunlight, harsh environmental conditions, or emotional states.
The team says: "Exposure to such face-altering factors, in turn, is associated with psychological traits. Liberals, for example, tend to smile more intensely and genuinely, which, over time, leaves traces in wrinkle patterns.
"Conservatives tend to be more self-disciplined and are thus healthier, consume less alcohol and tobacco, and have a better diet, altering their facial fat distribution and skin health"
The team says that the analysis of facial features associated with political orientation revealed that conservatives tended to have larger lower faces.
They add: "Moreover, the associations between facial appearance and political orientation seem to generalize beyond our sample: The predictive model derived from standardized images (while controlling for age, gender, and ethnicity) could predict political orientation from naturalistic images of 3,401 politicians from the United States, the United Kingdom, and Canada.
In conclusion, the team says: "Perhaps most crucially, our findings suggest that widespread biometric surveillance technologies are more threatening than previously thought.
Originally published by Talker News