A breakthrough study that revealed artificial intelligence could accurately guess whether a person is gay or straight based on their face is fast becoming viral all over the world.
Using pictures from a dating website with public profiles, researchers from Stanford University analyzed 130,741 images of 36,630 men and 170,360 images of 38,593 women. Using a computer algorithm, they found artificial intelligence was able to distinguish between gay and straight men with 81 percent accuracy and 74 percent for females.
Comparatively, human judges were only able to distinguish with 61 per cent accuracy for men and 54 per cent for women.
"[This means] faces contain much more information about sexual orientation than can be perceived and interpreted by the human brain," the authors wrote, reports The Guardian.
According to the study, the software was able to distinguish sexual orientation by picking up on subtle differences in facial structure.
The data found homosexual men had narrower jaws, larger foreheads and longer noses, while gay women had larger jaws and smaller foreheads than their straight counterparts.
While impressive, researchers admit the study had limitations as there was no consideration of transgender or bisexual people.
The author also preempted several criticisms by stating that this type of technology already exists and the purpose of his research was to expose security flaws and develop protections so that someone couldn’t use it for ill will.
"One of my obligations as a scientist is that if I know something that can potentially protect people from falling prey to such risks, I should publish it," Michal Kosinksi, co-author of the study, told The Guardian. He added that discrediting his research wouldn’t help protect LGBTQ people from the potentially life-threatening implications this kind of technology has.
The study, which was published in the Journal of Personality and Social Psychology, did not develop its own AI for the research and only tested existing technology.
Using pictures from a dating website with public profiles, researchers from Stanford University analyzed 130,741 images of 36,630 men and 170,360 images of 38,593 women. Using a computer algorithm, they found artificial intelligence was able to distinguish between gay and straight men with 81 percent accuracy and 74 percent for females.
Comparatively, human judges were only able to distinguish with 61 per cent accuracy for men and 54 per cent for women.
"[This means] faces contain much more information about sexual orientation than can be perceived and interpreted by the human brain," the authors wrote, reports The Guardian.
According to the study, the software was able to distinguish sexual orientation by picking up on subtle differences in facial structure.
The data found homosexual men had narrower jaws, larger foreheads and longer noses, while gay women had larger jaws and smaller foreheads than their straight counterparts.
While impressive, researchers admit the study had limitations as there was no consideration of transgender or bisexual people.
The author also preempted several criticisms by stating that this type of technology already exists and the purpose of his research was to expose security flaws and develop protections so that someone couldn’t use it for ill will.
"One of my obligations as a scientist is that if I know something that can potentially protect people from falling prey to such risks, I should publish it," Michal Kosinksi, co-author of the study, told The Guardian. He added that discrediting his research wouldn’t help protect LGBTQ people from the potentially life-threatening implications this kind of technology has.
The study, which was published in the Journal of Personality and Social Psychology, did not develop its own AI for the research and only tested existing technology.
No comments:
Post a Comment