• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

New AI can guess whether you're gay or straight from a photograph, 91% of the time

The study from Stanford University – which found that a computer algorithm could correctly distinguish between gay and straight men 81% of the time, and 74% for women – 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 research found that gay men and women tended to have ”gender-atypical" features, expressions and ”grooming styles", essentially meaning gay men appeared more feminine and vice versa. The data also identified certain trends, including that gay men had narrower jaws, longer noses and larger foreheads than straight men, and that gay women had larger jaws and smaller foreheads compared to straight women.

Human judges performed much worse than the algorithm, accurately identifying orientation only 61% of the time for men and 54% for women. When the software reviewed five images per person, it was even more successful – 91% of the time with men and 83% with women. Broadly, that means ”faces contain much more information about sexual orientation than can be perceived and interpreted by the human brain", the authors wrote.

The paper suggested that the findings provide ”strong support" for the theory that sexual orientation stems from exposure to certain hormones before birth, meaning people are born gay and being queer is not a choice. The machine's lower success rate for women also could support the notion that female sexual orientation is more fluid.

While the findings have clear limits when it comes to gender and sexuality – people of color were not included in the study, and there was no consideration of transgender or bisexual people



https://www.theguardian.com/technol...ether-youre-gay-or-straight-from-a-photograph

I'm not even sure what to think about this. I didn't think gay people looked a certain way.
 

Phased

Member
So.. the "gay radar" thing is real?

always thought its a superfluous BS

I think it's more about body language some people are good at picking up on rather than studying facial features, usually girls are pretty good at it in my experience.
 

Kurdel

Banned
1388702605496.gif
 

Fuchsdh

Member
So.. the "gay radar" thing is real?

always thought its a superfluous BS

Yeah? Any socially aware person can pick better-than-random-odds about this stuff. The fact that they can get a computer to a much higher frequency is interesting, though.

(And the OP's mistake is an easy way to see who actually read the article or excerpts—it's only 81% for gay men. The higher number comes with larger datasets.)

Perhaps the most interesting nugget is that it was only done on a caucasian sample. Would the comparatively wider number of facial features from more disparate ethnic groups confound the algorithms? Or would their physical observations hold true across phenotypes?
 

kadotsu

Banned
For the cross validation did they split the dataset on the personal data or image level. Some subjects had multiple images. I think they split it over persons so the NN didn't just learn face recognition.

Still very interesting. Unfortunately NNs don't tell us how the classification was made. I hope they can reproduce the classification with something less blackbox.
 

lenovox1

Member
So.. the "gay radar" thing is real?

always thought its a superfluous BS

Creating this tech is the definition of just because you can doesn't mean you should

This is something that's been researched for a while by a variety of research facilities.

The role of facial masculinity/femininity in the attribution of homosexuality (Aug 1990)
Facial Stereotypes of Deviants and Judgments of Guilt or Innocence
Physical Attractiveness Stereotype and the Attribution of Homosexuality Revisited
Perceived Sexual Orientation Based on Vocal and Facial Stimuli Is Linked to Self-Rated Sexual Orientation in Czech Men

Et al.


It seems to be the primary way queer and non-queer people identify other queer people, so it's important to research further and further...
 

FlyinJ

Douchebag. Yes, me.
Ugh, i can already sense how the homophobes will spin this shit to discriminate more.

On the flip side, homophobes will be outted as gay as well.

I think it may be an overall net positive, as 92% of the worst and loudest homophobes are closeted.
 

Fuchsdh

Member
What if someone is bi?

As the article points out, that's one of the limitations. Bi and trans/otherwise genderqueer people weren't studied for this purpose. If I had to hazard a guess, though, I'd bet they could get it working with a reduced but still better-than-human accuracy rate with those groups (it makes a certain inherent sense that bi people would still be more likely than the straight population to follow the markers they found in this analysis.)
 

Laughing Banana

Weeping Pickle
The idea that people with certain sexuality has physical characteristics that others can use to distinguish it seems... crazy to me.

And maybe demeaning as well? It feels like it strengthens the "they are the other" sentiment for me. It's weird.
 

lenovox1

Member
So erm how is this useful to humanity?

How much of the billions of things humans research on a daily basis meets your useful to humanity scale?

But it corroborates others' results on this subject and attempts to link biological factors to human sexuality.
 

Switch Back 9

a lot of my threads involve me fucking up somehow. Perhaps I'm a moron?
My sister's very first words after being told about this:

"This is stupid. What about Bears?"
 
Yeah Lesley Stahl explored this on 60 minutes about a decade ago, where she was asked to guess whether or not a man was gay or straight based on his silhouette during an interview; she guessed correctly most of the time. I think if you remove the rhetorical straightjacket of considering the idea that gay men or women may in aggregate look or sound identifiably "gay" as problematic, it's an interesting way to see how science and society interact.

The 60 Minutes segment also explored reasons one twin might be gay and the other straight, but the science on the matter is still young.
 
As the article points out, that's one of the limitations. Bi and trans/otherwise genderqueer people weren't studied for this purpose. If I had to hazard a guess, though, I'd bet they could get it working with a reduced but still better-than-human accuracy rate with those groups (it makes a certain inherent sense that bi people would still be more likely than the straight population to follow the markers they found in this analysis.)
I see, thanks.
 

III-V

Member
all the computer can do is recognize certain common traits and make a determination based upon probability. Not much more than that.

As long as this is not pseudo science or a flat out lie of a publication, I am sure there are better uses for the technology.
 

Air

Banned
Given enough exposure I think you can develop a good gaydar. Mines is pretty good but I don't know if I could tell through a photo.

I wonder how I'd show up. I'm straight but do have more effeminate features and have been mistaken for gay before
 

lenovox1

Member
My sister's very first words after being told about this:

"This is stupid. What about Bears?"

A beard is one masculine feature out of many facial traits that would have been used in their data set.

Unless she thinks all bears look straight, and that's just a personal "miss" on her part.
 

Laughing Banana

Weeping Pickle
the traits may not be culturally homogenous.

Like they even care. If someday a method is found that allows the existence of say, "Gay Detecting Device" or something, do you think they would care about "non-homogenous traits" when they point that device to person A and the device tell them that person A is gay?

Most probably they would lash/kill that person, uncaring about his/her defense.
 
It's actually the perfect test for homophobes, anti-gay lawmakers, priests, etc.

Denial is a common thing, what are you gaining by outing the gay homophobes?

And i really don't think we can make this kind of generalizations, as a poster above said, what about those who don't fall into the "feminine" category? The gay people i know are all different and not everyone has the same mannerisms. This is phrenology level shit.
 

We can research shit without creating essentially an gay detection machine.
 
This is really dumb and pointless

Sexuality isn't a look or a way of life it's just what kind of person you like. Even if this works what actual purpose does this thing serve.
 
Top Bottom