AI Can Tell If You're Gay: Artificial Intelligence Predicts Sexuality From One Photo with Startling Accuracy

Artificial intelligence can now tell whether you are gay or straight simply by analyzing a picture of your face.

Two Stanford University researchers have reported startling accuracy in predicting sexual orientation using computer technology.

Dr. Michal Kosinski and Yilun Wang, whose research will be published by the Journal of Personality and Social Psychology, say that AI can distinguish between the face of a heterosexual man and a homosexual man in 81 percent of cases. For women, the predictive accuracy is 71 percent.

In comparison, the scientists say, the average human is less adept at identifying between straight and gay people purely based on an image: We are only able to guess correctly in 61 percent of cases for men and 54 percent for women.

Facial recognition software A man uses a portable iris recognition scanner during the Biometrics 2004 exhibition and conference in London on October 14, 2004. Researchers found that AI could detect people's sexuality from a single photo with startling accuracy. Ian Waldie/Getty

When scientists presented the algorithm with five facial images of a single person, the accuracy increased to 91 percent for men and 83 percent for women.

Kosinski and Wang used “deep neural networks” to sample 35,326 facial images of men and women taken from a dating website.

Advanced computer analysis compared different facial characteristics and found that gay men and women tended to have “gender-atypical” features: fixed features such as the shape of one’s nose or jaw, and transient features such as hairstyle and facial hair.

The scientists then tested this idea to measure the "facial femininity" of each image, or the probability of that face being female. “The results show that the faces of gay men were more feminine and the faces of lesbians were more masculine than those of their respective heterosexual counterparts,” the study said. “Among men, the data-driven measure of facial femininity positively correlated with the probability of being gay.”

The findings advance discussion about the biological factors that may determine one’s sexual orientation. However, Kosinski tells The Economist, the research is not intended to be used to profile or “out” homosexual men and women. Rather, it is designed to demonstrate—or even warn—that technological advances can be used for such means and could pose a threat to our privacy, given that digital information is so easily accessible.

The researchers argue the “digitalization of our lives and rapid progress in AI continues to erode the privacy of sexual orientation and other intimate traits.”

Join the Discussion