Conservative politicians look like happier and more attractive in photos than liberal politicians, based on a latest study by artificial intelligence in Denmark.
Published within the journal Scientific Reports owned by Nature in March, research showed that AI could predict an individual’s political ideology with 61% accuracy by analyzing only one headshot.
Researchers fed roughly 3,200 publicly uploaded photos of political candidates who ran within the 2017 Danish municipal elections into Microsoft’s Azure Face API to evaluate an individual’s emotional state.
The evaluation showed that 80% of the faces had a joyful expression and 19% were neutral.
“Women (though not men) found high attractiveness scores amongst those identified by the model as more likely to be conservative,” the outcomes read. “These results are plausible on condition that previous human evaluator studies have also highlighted the link between attractiveness and conservatism.”
The results were much more accurate for men, 65%, before the researchers removed visual images aside from the person’s face – corresponding to shirt collars – from their photos.
The study found that left-wing male politicians had more neutral, less cheerful faces than their conservative counterparts.
“Attraction was not the one correlate of the ideology predicted by the model,” the researchers explain. “We also found that expressing happiness is related to conservatism for each sexes.
“Previous work has recognized that smiling in photos is a very important indicator of extroversion,” they continued. “And while extroversion will not be widely related to ideology, some studies have found that right-wing politicians are more extroverted.”
The researchers noted that “since attractiveness generally aids electoral success, all candidates are encouraged to submit a sexy photo.”
“Left and right-wing politicians could have different motivations for smiling – for instance, smiling faces have been found to look more attractive, which is comparatively essential for conservative politicians,” the paper reads.
“Future work is required to explore the extent to which joyful faces indicate conservatism outside of political samples.”
Of greater concern is the “privacy threat posed by deep learning approaches” using publicly available data.
This is not the primary time AI has raised warning flags.
In March, Facebook removed artificial intelligence-generated fake sex ads on social media that used the likenesses of actresses Scarlett Johansson and Emma Watson.
And a recent report found that generative AI will cause significant disruption to jobs occupied by “higher-paying knowledge employees,” whose roles were “previously considered relatively proof against automation.”
