Conservative women are more attractive than liberals, study says

Conservative women are more attractive than left-wing women, according to a European survey of thousands of faces.

Danish and Swedish researchers tested a deep-learning artificial intelligence called a neural network that can usually predict a person’s political leanings based solely on their portrait photo.

It showed that right-wing women were more attractive, based on a publicly available scoring system. The group found no such link in men, but did find that the left-wing men showed more neutral, less happy faces, suggesting they may be better able to guard their emotions.

However, the real goal of the researchers’ study was to demonstrate the alarming accuracy of out-of-the-box AI, which can correctly guess a person’s political views based on limited information, such as a simple selfie taken every day on social media is posted.

“Our results confirmed the threat to privacy from deep learning approaches,” the researchers wrote in their published findings. the Nature magazine Scientific Reports.

The researchers, a trio of psychologists and political scientists from Denmark and Sweden, selected 3,323 publicly submitted portrait photos of political candidates for analysis in a neural network capable of encoding and classifying facial expressions.

The photos all came from candidates running in the 2017 Danish municipal elections, submitted by the candidates themselves to the Danish Broadcasting Cooperative.

For both men and women, the AI ​​was correct at predicting a person’s political leanings based on just one photo of their face 61 percent of the time.

The results were even more accurate for men, 65 percent, before the team stripped their photo dataset of all visual images other than the man’s face.

Given the low level of political polarization in Denmark and lower political interests in these municipal races, the researchers theorized that these candidates most likely reflected the faces of ordinary, everyday, politically partisan individuals.

Danish political scientists have affectionately dubbed these local candidates the “last amateurs in politics,” making them ideal surrogates for testing how AI might succeed in guessing the average person’s politics — from just a few photos online.

Danish researchers have tested a neural network that can predict someone’s political leanings based solely on their portrait photo. Heatmaps (pictured) revealed which parts of the women’s and men’s faces the neural net targeted to make its eerily accurate political predictions

While the AI ​​found that women with more attractive features were more likely to be politically conservative, it found no such connection when analyzing its pool of photos of male politicians. For both men and women, the AI ​​was correct 61 percent of the time

The researchers dropped 188 individuals from the original dataset based on their non-European ethnicity, saying those candidates were more than 2.5 times more likely to represent a left-wing party and to have racially biased the AI.

Face API from Microsoft’s Azure Cognitive Services was used by the neural network to assess the emotional state expressed in each candidate’s photo. The results determined that 80 percent of the faces showed a happy expression, while 19 percent read as neutral.

Alarmingly, the AI ​​was even more accurate for men, at 65 percent, before the team stripped their dataset’s photos of visual images other than the man’s face (as pictured above).

“Using a pre-developed and readily available network trained and validated solely on publicly available data,” they wrote in their conclusion, “we were able to identify the ideology of the person depicted about 60% of the time. to predict.’

While the Danish team determined that this was partly due to bias in the type of photos politicians presented to the public and partly due to Face API having difficulty identifying certain other facial expressions.

However, the deep-learning algorithm caught one unusual distinction: Female, left-wing political candidates were slightly more likely to be read by the neural net as a face expressing contempt, the researchers said.

The team was much more confident in the neural net’s findings, linking a high attractiveness score to conservative views.

“These results are credible given that previous research with human raters has also shown an association between attractiveness and conservatism,” the authors wrote.

While the deep learning algorithm found attractiveness to be a predictor of political leanings for women, the same couldn’t be said for male candidates.

Neither could macho stereotypes: The Neural Net, based on its programmed indicators of male facial features, could not detect noticeable differences between right- and left-handed men.