AI finds conservative women more attractive and happy in photos.

Fashion and beauty

as if Jack Hobbs

June 20, 2023 | 5:28 p.m

Conservative female politicians appear happier and more attractive in photos than liberal politicians, according to a new artificial intelligence study conducted in Denmark.

Research published in March in the proprietary journal Scientific Reports found that AI could predict a person’s political ideology with 61% accuracy by analyzing just one head.

The scientists fed photos of 3,200 political candidates who ran in the 2017 Danish municipal elections into Microsoft Azure’s Face API tool to assess the individual’s emotional state.

The analysis showed that 80% of the faces showed a happy expression and 19% were read as neutral.

“Higher beauty scores were found for women (but not men) among those identified as potentially conservative by the model,” the findings read. “These results are credible because previous studies using human measurements have shown a relationship between attractiveness and guarding.”

The results were 65% more accurate for men, before the researchers removed other visual images from the man’s face – such as shirt collars.

Left-leaning male politicians show more neutral and unhappy faces than their conservative counterparts, the study found.

“Ideological attractiveness was not the only correlation predicted by the model,” the scientists explained. “We also found that expressing happiness was associated with conservatism for both sexes.

“Previous work has found smiling in photographs to be a valid predictor of outcome,” they continued. “Although not broadly related to ideology, some studies have found that right-wing politicians are more marginalized.

“All candidates are encouraged to submit an attractive photograph because attractiveness generally contributes to electoral success,” the scientists said.

“Politicians on the left and right may have different motivations for smiling – for example, smiling faces are found to be more attractive, which is relatively beneficial for conservative politicians,” the paper reads.

“Future work is needed to examine the extent to which happy faces indicate conservatism outside of samples of politicians.”

More concerned is the “privacy risk posed by deep learning approaches” using publicly available data.

This isn’t the first time AI has raised warning flags.

In March, Facebook removed AI-generated deep fake sexual social media ads featuring avatars of actresses Scarlett Johansson and Emma Watson.

And generative AI will cause significant disruption to jobs held by “high-wage knowledge workers” that were “previously considered relatively immune to automation,” according to a recent report.