An artificial intelligence algorithm can allegedly predict a person’s political beliefs based on photographs of their face, at least according to a new study.
The study, conducted in Denmark, claimed that people who smile in photographs are more likely to hold right-wing views, while those with neutral expressions are more likely to hold left-wing political beliefs.
The scientists conducted their study using “deep learning” technology — a form of AI that trains computers to process data in a fashion modeled after the human brain.
Deep learning can allegedly be used to predict various characteristics of people based solely on photographs of their faces. The researchers found that through deep learning, AI can analyze a photograph of an individual and predict their political leanings with 61% accuracy.
Still, the Danish researchers — Stig Hebbelstrup Rye Rasmussen, Steven G. Ludeke, and Robert Klemmensen — claimed the technology should raise concerns about personal privacy.
“Our results confirmed the threat to privacy posed by deep learning approaches,” the researchers wrote, adding that the technology could affect employment opportunities.
“The social importance of the question is clear from a privacy protection perspective: Facial photographs are commonly available to potential employers, and those involved in hiring decisions self-declare a willingness to discriminate based on ideology,” the study reads.
In a study published last year, researchers were able to use AI to find “noticeable and discriminative [facial] features among liberals and conservatives.”
Researchers concluded that their study “suggests that the biological and neurological roots of political behavior run much deeper than previously thought.”
The Danish study also found using deep learning AI that “right-wing faces, particularly for females, might be perceived as more attractive.” A prior study similarly found that right-wing politicians are, in general, considered more attractive than those who are liberal.
The study analyzed over 4,600 politicians in Danish municipal elections from 2017, most of whom were male. The researchers excluded subjects who “could not be readily assigned an ideological score” or were not of European ethnic origin.