Technology » Science

Gay Mirror: New AI Can Determine Your Sexuality from a Single Photo

Friday Sep 8, 2017

No, this isn't the plot of an episode of "Black Mirror."

New artificial intelligence out of Stanford University can reportedly accurately figure out if a person is gay or straight - based on a single photo, the Guardian reports.

According to new research, the algorithm from the college can correctly distinguish between gay and straight men 81% of the time and 74% of the time for women. Not only does the new study suggest that computers have better "gaydar" than humans but the Guardian writes the new A.I. "has raised questions about the biological origins of sexual orientation, the ethics of facial-detection technology and the potential for this kind of software to violate people's privacy or be abused for anti-LGBT purposes."

The algorithm was published in the Journal of Personality and Social Psychology and first reported on by the Economist.

The study was based on a sample of more than 35,000 facial images from men and women publicly posted on a dating website. Researchers Michal Kosinski and Yilun Wang said they "extracted features from the images using 'deep neural networks,' meaning a sophisticated mathematical system that learns to analyze visuals based on a large dataset," the Guardian writes.

The researchers said gay men and women tend to have "gender-atypical" features, expressions and "grooming styles." This basically means gay men appeared more feminine and gay women more masculine. The data also found trends among gay men: They tended to have narrower jaws, longer noses and larger foreheads compared to straight men. Gay women apparently had larger jaws and smaller forehands compared to straight women.

It was reported humans were significantly worse at trying to identify if someone was gay or straight compared to the algorithm, which correctly identified the orientation 61% of the time for men and 54% for women. When the A.I. looked at five images of a single person, it's success rate was even higher with correctly choosing a person's orientation 91% of the time for men and 83% of the time for women.

This means "faces contain much more information about sexual orientation than can be perceived and interpreted by the human brain," the researchers wrote. The report suggests the data provide "strong support" for the theory that sexual orientation stems from exposure to certain hormones before birth - meaning people are born gay. That the algorithm had a lower success rate for women also suggests that female orientation could be more fluid.

The Guardian points out people of color were not included in the study nor were transgender or bisexual individuals. Nevertheless, the newspaper writes the "implications for the artificial intelligence are vast and alarming."

"With billions of facial images of people stored on social media sites and in government databases, the researchers suggested that public data could be used to detect people's sexual orientation without their consent," the Guardian writes.

"It's certainly unsettling. Like any new tool, if it gets into the wrong hands, it can be used for ill purposes," Nick Rule, an associate professor of psychology at the University of Toronto, who has published research on the science of "gaydar" told the newspaper. "If you can start profiling people based on their appearance, then identifying them and doing horrible things to them, that's really bad."

Click here to read the study and click here to read the Guardian's full report on the study.


Add New Comment

Comments on Facebook