Slashdot Mirror
AI Can Detect Sexual Orientation Based On Person's Photo (cnbc.com)
ugen shares a report from CNBC: Artificial Intelligence (AI) can now accurately identify a person's sexual orientation by analyzing photos of their face, according to new research. The Stanford University study, which is set to be published in the Journal of Personality and Social Psychology and was first reported in The Economist, found that machines had a far superior "gaydar" when compared to humans. Slashdot reader randomlygeneratename adds: Researchers built classifiers trained on photos from dating websites to predict the sexual orientation of users. The best classifier used logistic regression over features extracted from a VGG-Face conv-net. The latter was done to prevent overfitting to background, non-facial information. Classical facial feature extraction also worked with a slight drop in accuracy. From multiple photos, they achieved an accuracy of 91% for men and 83% for women (and 81% / 71% for a single photo). Humans were only able to get 61% and 54%, respectively. One caveat is the paper mentions it only used Caucasian faces. The paper went on to discuss how this capability can be an invasion of privacy, and conjectured that other types of personal information might be detectable from photos. The source paper can be found here.
Maybe they're getting interference from a gay weather balloon.
#DeleteChrome
...definitely applies to this situation. This has some pretty negative implications in particularly homophobic regions. All the more reason not to visit the pacific northwestern US or the middle eastern region in general if this thing gets to be widespread.
Furries make the internet go.
If the AI shows me a picture of Super Pochaco, can it reliably detect my preference? Or maybe it needs a picture of something else on my body?
Mostly random stuff.
If sexual orientation correlates highly with physical appearance, then I think this conclusively proves that sexual orientation is not a "decision."
* Unless this is picking up on subtle cues like gay men wearing eyeliner and gay women not wearing makeup. (Similar to lots of how gay men speak with an "affliction" and drive Saabs, while gay women drive Subarus.)
Is the sample data biased toward gay individuals or something? 91% and 83% just doesn't seem that impressive compared with the error rate you would get by just "guessing" straight every time for the general population, and 61% and 54% are way off.
Or David Duke or Rick Santorum or "Pastor" Jack Graham or Rick Perry. Them Texas boys are all hidin' something. Mitch McConnel and "Randy" Paul. How about Orrin Hatch?
The Russians have won. They have made the world a cesspool of distrust, greed, fear and hate.
I would LOVE to get a copy of this program and, not only try it out on myself and my friends, (I think we're relatively secure in our sexuality) but try it on famous people.
Specifically: Republican lawmakers and perhaps even Christian preachers! (How about Mr. Macho himself, Putin?)
I really, really know I'm going to be down-modded for this but please hear me out. Haven't you wondered why those people who are so against homosexuality often turn out to be gay themselves? (Dennis Hastert and that lawmaker caught in the men's restroom soliciting a cop come to mind). Maybe it's because they are so ashamed that the only way they can bury their feelings is to actively suppress it. That's fine if you don't want to face the truth but the problem is being lawmakers, representatives of God, they infringe on many, many other peoples lives. So let's drag them out of the closet and into the photo booth! (Actually I don't think that'll be necessary, from what little I've read about this algorithm it doesn't require any particular lighting or "orientation" (ha ha) for the photo so many of the pictures of these famous people should be just fine.)
On a more serious note: This is just the latest in a trend of events which a friend of mine has said is "the end of privacy". With technologies like these (soon I'm sure they'll be able to analyze videos to see, by looking at imperceptible* subtle face color flushing and breathing patterns, who is attracted to whom), social media and the hack of personal databases like Equifax, NOTHING will be able to be kept secret. I wouldn't doubt that the CIA is already using some of this stuff to determine, remotely, if someone is lying on camera when they say something. It will be hard to legislature laws to keep it out of business and impossible to keep out of statecraft.
*imperceptible to humans
Ive always said the danger with AI isn't killer robots but killer humans. Machine learning is being used to perpetrate a huge invasion of privacy in the form of "big data" data matching. It's like countless companies , and governments, have deployed armies of robot detectives to sift out or repeat secrets , and not to solve crimes or whatever but to manipulate us into compliant consumers. This particularly feat is even more worrying however because I'm certain theres any number of theocratic fascists regimes , Christian , Muslim and beyond who would be very interested in this. Gay pre-crime , so to speak. Welcome to the future
Excuse the Unicode crap in my posts. That's an apostrophe, and slashdot is busted.
"The paper went on to discuss how this capability can be an invasion of privacy, and conjectured that other types of personal information might be detectable from photos."
Any information that can be gleaned from a public photo is public, and has no expectation of privacy.
I always guess straight, and I'm right ~96% of the time!
As usual, the summery isn't accurate. Accuracy at determining sexual orientation is not what was measured. They measured the ability when given two photos, one of a gay person, and the other straight to pick which is which.
Also, the photos were from dating sites and showing people who likely put in an effort to appear they way they classified themselves.
So really, they have an AI that can tell the difference between gay dating profile pictures and straight ones when given multiple images, and a single gay and a single straight candidate to pick from. I honestly wouldn't be surprised if the lighting and pose and camera positioning were enough to do that well.
Overall its an impressive result, but it doesn't mean you will be able to point your phone at someone to know if they are gay any time soon. It is a suggestion that we may get there eventually though.
The paper went on to discuss how this capability can be an invasion of privacy, and conjectured that other types of personal information might be detectable from photos.
While walking about yesterday I remember seeing two people that stuck out to me. The dude in the parking lot with the wild hair, wearing eyeliner, a tight black crop top, and purple pants? Probably gay. The short lady that I shared an elevator with, wearing brown jeans, a buttoned up long sleeve shirt, short hair, and comfortable shoes? Probably gay.
We say a lot about ourselves with how we dress and act. All it takes is someone to notice. There are people trained in this for lots of reasons. I remember seeing a note in my medical records that said something like, "patient was well dressed, and well groomed." I thought that was something odd for a physician to make a note of in a medical record. That is until I thought about it some more. The guy wasn't just evaluating my physical health but my mental health. Someone that shows up for an annual physical exam with bed head, still wearing the clothes they slept in, and stinking like they haven't bathed in days could mean they have mental problems, mobility problems, etc.
Even people not formally trained in looking at such details will notice these things. I'll go to a hardware store wearing some new jeans (carpenter style even, with hammer loops), a button up shirt, and clean running shoes and I get treated one way. I go the next day wearing paint stained jeans (also carpenter style), a faded t-shirt, and work boots and I get treated differently. I've learned to dress for the part I'm playing that day.
There's no "invasion of privacy" here. People are advertising who they are everyday.
I am armed because I am free. I am free because I am armed.
If the AI were to simply assign scores of "Straight" to EVERYONE, it would achieve 90% accuracy for men and 85% accuracy for women, since about 90% of men are straight and about 85% of women are straight. So scores of 91% and 85% accuracy are not statistically significant.
My kid bro gave Mom a Subaru...what has he done?
It's an invasion of privacy? Bro, you're broadcasting it. Just stop looking so gay. This is not a serious post.
Gay people can distinguish other gay people with MUCH greater frequency than straight people can.
I have no doubt. After being in the Army I can "smell" military experience on people. People in general can see things in others that they have experienced themselves.
Little things can tell people a lot. I learned from living in Texas that the plural of "you" is not "y'all". The proper way to address a group is "all y'all". The people that don't get that right are not from Texas. Maybe they are from Arizona, or Mississippi, I don't know because I haven't been to those states.
I am armed because I am free. I am free because I am armed.
This technology should be destroyed and the research should be buried. It could easily end up in Middle Eastern countries where they would use it to kill people based on the algorithm. Usually I don't support suppressing technology, but this is seriously a bad idea.
People don't even have 91% accuracy judging themselves on this.
... they finally got their fruit machine...
Run his face in that gay detecting AI (ANN?)
This is the sorting hat for a future dystopian Harry Potter remake.
AI can, in principle, do anything a computing machine can do in principle. As for sexual orientation, there is yet to be devised a sensible theory and classification which doesn't massively oversimplify human sexuality well past the point of taking the piss.
As such, they're playing the game of keeping definitions fuzzy enough that they can wing it when they write up the paper.
John_Chalisque
When I see ridiculous nonsense like this (just read the title and don't need anything else), the first thing coming to my mind is the surprisingly big number of clueless individuals, completely unaware about that fact, with stupid fears, concerns, expectations, etc. and usually affecting others in the most stupidly arbitrary and unfair way possible. People whose behaviours can be defined with "I don't know about this, but here you have my opinion", "if I see X, it would mean Y, otherwise Z; 100% accurate; will always trust it", "I am not being paid for knowing, but for making decisions", "this cannot be true, otherwise it would have already been done", etc.
Custom Solvers 2.0 = Alvaro Carballo Garcia = varocarbas.
Sometimes AI can do a lot of damage. In many countries homosexual people are prosecuted and cruelly executed. I can’t understand why someone at Stanford thought it was a good to write this.
"this capability can be an invasion of privacy"
As such, licensing fees instead of starting in 5 digits, now start in 6 digits.
Not only is gender determined biologically, gender identity can be detected with biological markers.
If a computer told a person the gender pronoun that their sexual preference aligned with, it would be the punchline of a joke so ironic you would have to take it as evidence there a god.
My ism, it's full of beliefs.
First off, what they claim to have created is tantamount to computer-assisted phrenology -- long since debunked and tossed on the scrapheap of superstition.
The most obvious flaw appears here, starting on line 208:
*headdesk*
Leaving aside the gigantic issues presented by self-reporting and self-selecting samples, these idiots failed to account for a common practice among hetero women on dating sites, which is to falsely claim to be seeking other women as a means to reduce or eliminate an onslaught of tacky propositions from clueless het-boys.
Other glaring flaws include:
An actual sociologist could probably identify dozens of other flaws, any one of which would be fatal to the work.
I would undertake to create a similar piece of software that tries to identify criminals from photographs, and use police mugshots to train it. Surprise! Black people are more likely to be criminals! GIGO.
Frankly, I think they should have taken their theme from the closing paragraphs of their paper: "We created a digital phrenologist out of deep neural networks and other off-the-shelf parts that coughs up results that seem relevant and meaningful to the layman, when in fact they're utter garbage." That would have been a good paper.
Perhaps we can indeed learn new things by letting a DNN stare at human faces. But IMHO this paper is utterly valueless in identifying what those might be. GIGO.
Editor, A1-AAA AmeriCaptions
Als like their smooth and smoky Scotch in a quiet bar.
Not all of them. I once met an AI which really enjoyed electronic music and tap water. Its significant other was a quite outgoing toaster with a weirdly deep knowledge about fly-crocodile mating. Curiously, I firstly met that colorful couple in a let's-ban-all-AI meeting organised by a hating-itself robot which I brought to life just for fun. LOL.
Custom Solvers 2.0 = Alvaro Carballo Garcia = varocarbas.
... which are detected as gay but are straight are just lying to themselves
About 10% of all people are homosexual (very roughly). Hence a static classifier of value="hetero" has an accuracy rating of around 90%. Given that, the published numbers are not impressive at all.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
I learned from living in Texas that the plural of "you" is not "y'all". The proper way to address a group is "all y'all". The people that don't get that right are not from Texas.
Actually, saying "all y'all" is incorrect, and the rest of the South laughs at Texas for getting it so wrong.
The current theory is prenatal hormones. The same ones that cause actual gender differentiation are also currently believed to effect sexual orientation later in life, which as well as being backed by science is also logical.
But we all know that the social constructionists abandoned this science the instant gay marriage was approved... so here way are..
"His name was James Damore."
instagram will never be the same once facebook buys this software
It was only 60 years ago that being homosexual was outlawed in most "civilized western nations" and "treated" with drugs and electroshock. Society doesn't have a direction of continued permissiveness and can slip back at any time. Imagine what this ability would then be used for.
Got you an extra Mom?
Am I missing something? This kind of software has existed for years in Google's Play Store and provides nearly instantaneous results instantly after analyzing a picture or a fingerprint.
Anecdotal evidence is just that. I once was teaching a grad. level logic course and two fellows were in the military taking the course. About half-way through, they confronted me saying I must have been in the military once. Nope, that wasn't me. Vibrations are weird, just remember the Summer of Love.
Since when did being a common everyday douche earn a death sentence?
I think you're a dochue... so it's ok to kill you now?
> which is to falsely claim to be seeking other women as a means to reduce or eliminate an onslaught of tacky propositions from clueless het-boys
How the hell is that going to get them any men at all? I've never heard of this practice and don't believe it - please show evidence this is routinely done
> All subjects caucasian
As the study says
> No attempt made to account for (or even acknowledge the existence of)
So they used self-identified gay/straight to try to keep a clear 'signal' in the noise. Sounds sensible.
> No attempt made to account for economic status and history
So physiognomy is affected by wealth - is sexuality too?
> No attempt made to account for cosmetic surgery or other such treatments
Possibly your only fair point IMO. Your trashing of a study is on an even worse basis than the study itself, it appears.
I would hardly call heavy makeup for a man a 'subtle' clue...
Anyway, 91% accuracy is complete disaster. While there is a common feeling that 10% of population is gay, more realistic studies (like ones referenced at https://www.theguardian.com/po...), claim between 1.5% and 6%. Even taking highest percentage into account (one provided by pro-gay organization), of 6%, I can write simple gaydar app which will tell 'straight' 100% of time and it will be right 96% of time.
You cannot take a single measure with x% of error to gain meaningful information about things which occur x% of time. It can be a screening test, but not a final answer. Simple example is machine which detects some rare disease and is wrong only 1% of the time. If disease happens for 1 person in million, when machine says you are ill, you have only 0.1% chance of being actually ill and 99.9% of chances that machine was wrong. What you can later do, is to put these 10000 people for more expensive/detailed/invasive tests, but not to start treating them for that disease outright.
The algorithm was trained on pictures from dating sites, which isn't exactly a representative data set. A large portion of the pictures have been manipulated or at least carefully selected by the person who uploaded them. There are a portion of gays who try to signal their gayness through things like styling and mannerisms, which would be easy for an algorithm to pick up on. These types are going to be overrepresented in the data set.
... that makes guarantee that it is correct.
Which side is earring on? Left is right and right is wrong.
Have gnu, will travel.
Many gays adopt certain dress styles and behaviors as a matter of self preservation (not getting the shit beat out of them for making passes at the 'wrong' person). From TFA:
The Stanford University researchers found that gay men and women typically had "gender-atypical" features and expressions. While a person's "grooming style" also factored in to the computer algorithm, essentially suggesting gay women appeared more masculine and vice versa.
It really isn't inconceivable that AI could be trained to recognize these features.
Have gnu, will travel.
The training and test samples are probably a 1:1 mix.
That kind of argument shows that a test is useless for large scale screening, but the test still gives you "meaningful information", otherwise we wouldn't be using HIV tests.
That which can be seen is public, not private. It really is that simple. Feelings are another issue. Many people want to be covert for a host of reasons and feel that even a license plate on their car somehow robs them of their privacy. It is 100% nonsense. The very real question is whether humans can exist in an honest world. I suspect that they can not. there is so much secretive nonsense in so may peoples' lives that society might become completely unhinged if we really know each other.
Social psychology is essentially a pseudo-science. I don't expect future studies to confirm this at all.
The Uncoveror: It's the real news.
Rightm it's not a "false categorization"... sexual orientation is real. Or have people been fighting all this time over nothing?
Fighting over the definition, not the groups gathering on each of the "voting points" i.e. Male-Female, Male-Male, Female-Male, Female-Female.
If you are a Female, looking for a Male, do you really need this? A gay man wont date you. No shit. No need for a "facial scan" to tell me to avoid someone who is going to say NO anyway.
The fuck does an invasion of privacy have to do with anything else, other than that. If you want to date someone, ask them out. If they say no, is it cause they are gay? Maybe, that is for them to let you know, not use a system of biased opinion to invade someone's privacy and assume.
And lets be real: Unless this thing is 100% accurate, why even bother? Any "bias" it injects is tantamount to social stereotyping, which, congrats morons! You just successfully programmed a COMPUTER TO DO SOCIAL STEREOTYPING. News at 11.
Cheers.
How is this fixing any sort of problems?
Is this just a party trick?
It seems like a pretty irrelevant issue. The moment you launch the software you are stamping a great big "B-I-G-O-T on your forward and losing any respect people had for you.
The northwest is tolerant. But you're a bunch of cunts.
We love you too, man.
Irony: Agile development has too much intertia to be abandoned now.
-1 Modded Truthful
If you are a Female, looking for a Male, do you really need this? A gay man wont date you. No shit. No need for a "facial scan" to tell me to avoid someone who is going to say NO anyway
No, but if you're a guy, looking for a guy, getting this wrong in the new deep south like Portland or Seattle will cost you teeth. Guess how I know?
Furries make the internet go.
This is poor science. The training set was from photos on a dating site from people who self-identified their sexuality. Presumably the photos weren't passport photos, but ones chosen by the individual to appeal to their targets. That's biased right there.
Also, there's an ethical question about using the photographs, as others have pointed out. I'm assuming that the subjects did not consent to have their images used for this purpose. They may be public images, but there's an ethical question about the use for such an experiment.
Is holding an Ig Nobel the goal? Get back to work.
If a picture can be used to determine sexuality, could a picture also be used to determine birth gender? If that came on the market just think of the LGBT?+ community would do!
...all the people who claimed they had a gaydar were right?
If they can only shrink it to a hand-held size, everyone can carry their own personal GayDar. And refine it to the point where it even classifies the level of the gay "Not gay, but has seen one up close". What a useful use of artificial intelligence. (/sarcasm)
There are 2 groups of people you can make fun of on the Internet without fear of attack. The illiterate, and the Amish.
...If they use this. Expect a whole lot of extremely angry and litigious individuals if they ever attempt to use this AI anhywhere.
http://how-old.net/
Casteism
Modded +1 Interesting! Pffff. Adding one "LOL" (= laughing out loud = I am joking) + answering to an evident joke by evidently extending the ideas in said joke is still too confusing for some people. And I am writing all this despite having fully accepted absolute (online) stupidity!
Custom Solvers 2.0 = Alvaro Carballo Garcia = varocarbas.
After seeing the last stupidity about LSD (to improve productivity!!), I do feel like writing a tiny clarification here to somehow account for the specific aspect of stupidity/ignorance which most likely justifies the aforementioned +1 interesting.
:)
Although I don't take any kind of drugs (including soft ones like alcohol or tobacco) since some years ago (never ever at work and I usually post on Slashdot from my office) and was never a heavy user neither tried many/hard ones, I did tried some drugs in the past mainly while being a student; always sporadically, when partying/having nothing relevant to do and by enjoying the experiences (= never felt anything even remotely close to any kind of addition). I did try some hallucinogens and did like them.
Despite my not-too-relevant experience with drugs, I think that my ideas about their most likely effects are quite clear. With hallucinogens, for example, you might get the point of seeing weird shapes/colours, even kind-of-imagining a small inexistent pseudo-reality, but mainly in that specific moment. When the effects disappear you should realise about pretty much everything, like waking up from a dream. Also your willingness to (not) see is also quite important. So, I don't think that it is possible that someone could ever think that some unrealistic craziness on drugs is real; unless that person wants to think/say that it is the case or under extreme conditions like a heavy user.
I am completely against any kind of drug in the work environment, simply because of being completely incompatible with what I consider work (= what I do): performing a demanding-enough engineering/programming work is pretty much the opposite to the relax-prone environment where certain substances are expected to be used. On the other hand, people performing artistic-/ideation-focused tasks might take some advantage from the stress-reduction which might be associated with the consumption of certain substance; but, objectively speaking, the capabilities of that person will be certainly reduced. Anyway, all that isn't applicable to me or to the kind of technical-intensive work environment in which I am looking to be involved; the only drugs which can affect my behaviour are illness, tiredness or similar
Custom Solvers 2.0 = Alvaro Carballo Garcia = varocarbas.