The thing that gives AI an upper edge over any other technology is it’s capability of recognizing the patterns and learning from past mistakes. It’s being now used to write poetry, spot micro symptoms of diseases and teach children’s about renewable energy, etc. Going a mile further, the researchers at Stanford University has created an artificial intelligence system that predict whether a person is gay or straight just by looking at his photo.
The creator of this AI is Michal Kosinski and Yilun Wang. The research’s research paper titled Deep neural networks are more accurate than humans at detecting sexual orientation from facial images, has published in Journal of Personality and Social Psychology. Researcher said ” faces contain much more information about sexual orientation than can be perceived and interpreted by the human brain”.
During research Michal Kosinski and Yilun Wang found that gay men and women tended to have “gender-atypical” features, expressions and “grooming styles”. Gay men were found to have narrower jaws, longer noses, and larger foreheads than their straight counterparts. Whereas a gay women were found to have larger jaws and smaller foreheads compared to straight women.
Lesbians tended to use less eye makeup, had darker hair, and wore less revealing clothes (note the higher neckline) indicating less feminine grooming and style. Furthermore, although women tend to smile more in general, lesbians smiled less than their heterosexual counterparts.
“The differences in femininity observed in this study were subtle, spread across many facial features, and apparent only when examining averaged images of many faces,” the researchers wrote in the paper.
For training this AI, researchers Michal Kosinski and Yilun Wang, extracted features from the images of more than 35,326 Americans from public profiles on a US dating website. The images then fed into a software named VGG-Face, which assigns a ‘faceprint’ number to each face. Researchers used between one and five of each person’s pictures and took people’s sexuality as self-reported.
The accuracy of this AI system is insanely high. It could correctly predict whether a man is a gay at an accuracy of 91 percent of the time and a woman 83 percent of the time. If we compare it to humans, accuracy of a human is too poor. A person can accurately identify only 61% of the time for men and 54% for women.
Kosinski said “Our results provide strong support for the [prenatal hormone theory], which argues that same-gender sexual orientation stems from the underexposure of male fetuses and overexposure of female fetuses to prenatal androgens responsible for the sexual differentiation of faces, preferences, and behavior.”
It also stressed that the findings don’t imply that “all gay men are more feminine than all heterosexual men, or that there are no gay men with extremely masculine facial features (and vice versa in the case of lesbians)”, and must not be misinterpreted.
This AI is still in it’s development stage and has some clear limits in terms of human gender and sexuality. In proposed study color of people has not include. It also does not mention transgender or bisexual people in the study.
“We were really disturbed by these results and spent much time considering whether they should be made public at all. We did not want to enable the very risks that we are warning against,” Kosinski and Wang wrote in author notes published along with the paper
Kosinski believes that his technology in future could be used to detect emotions, IQ and even a predisposition to commit certain crimes.
He said “We were really disturbed by these results and spent much time considering whether they should be made public at all. We did not want to enable the very risks that we are warning against
You can find this research paper in this PDF format.
More in AI :