A Better Gaydar (or "He Looks Gay")

[Oxx]

It would appear that computers can now identify a man's sexuality from photographs with more than 90% accuracy - far better than human beings.

 

Hard to believe, but apparently true.  Now, if they can link the processing to an app that would allow me to photograph a random stranger and let me know whether he would like me to shag him senseless get to know him better, dating could be revolutionised.

 

https://techcrunch.com/2017/09/07/ai-that-can-determine-a-persons-sexuality-from-photos-shows-the-dark-side-of-the-data-age/

[Jingthing]

It's not an issue if homosexuality was totally destigmatized. So it is an issue.

 

Anyway, it's almost impossible to stop technological "progress" from moving forward.

 

Back in the day, it was different ...

 

 

Edited September 8, 2017 by Jingthing

[canuckamuck]

Why do we need this, mostly you know pretty quick, and usually it makes no difference.

[Jingthing]

40 minutes ago, canuckamuck said:

Why do we need this, mostly you know pretty quick, and usually it makes no difference.

In many parts of the world, you can be killed for being gay. So it can make a big difference for those that need to be in the closet for the sake of survival. 

[canuckamuck]

Touché

[tominbkk]

3 hours ago, Jingthing said:

It's not an issue if homosexuality was totally destigmatized. So it is an issue.

 

Anyway, it's almost impossible to stop technological "progress" from moving forward.

 

Back in the day, it was different ...

 

 

 

[mogandave]

So if it is only 90% accurate, could it not just say everyone was heterosexual and claim to be within tolerance?

 

 

 

[Jingthing]

So if it is only 90% accurate, could it just not say everyone was heterosexual claim to be within tolerance?


Sent from my iPhone using Thaivisa Connect

Well an oppressive regime could do a quick screen with that tool and then investigate the positives to confirm. Such as anal exams such as are done in Iran and penis engorgement tests using different flavor of porn.

[onthesoi]

Lol, bullshit!

[mogandave]

Well an oppressive regime could do a quick screen with that tool and then investigate the positives to confirm. Such as anal exams such as are done in Iran and penis engorgement tests using different flavor of porn.

Or just wait for the guys that apply for jobs as screeners and engorgement testers...

[Jingthing]

Or just wait for the guys that apply for jobs as screeners and engorgement testers...

Might be fun in the context of a fetish club.

[mogandave]

Might be fun in the context of a fetish club.

Or a game show...

[onthemoon]

On 9/9/2017 at 7:58 PM, onthesoi said:

Lol, bullshit!

Actually, I do believe it is BS.

 

However, if enough people - or the important people - believe it to be true, it can become really dangerous. Dangerous in the sense that the Nazis would have loved to have a tool like this. 

[siam2007]

16 minutes ago, onthemoon said:

Actually, I do believe it is BS.

 

However, if enough people - or the important people - believe it to be true, it can become really dangerous. Dangerous in the sense that the Nazis would have loved to have a tool like this. 

 

Islamic governments, even here in South-East-Asia (Indonesia, Malaysia) may love that tool as well.

 

[mogandave]

Even if it worked, would it no be easy enough to get a list of the "tells" such that detection could easily be avoided?




Sent from my iPhone using Thaivisa Connect

[onthemoon]

6 hours ago, mogandave said:

Even if it worked, would it no be easy enough to get a list of the "tells" such that detection could easily be avoided?

It appears you didn't read the article. Let me quote this part for you:

 

Quote

Using a database of facial imagery (from a dating site that makes its data public), the researchers collected 35,326 images of 14,776 people, with (self-identified) gay and straight men and women all equally represented. Their facial features were extracted and quantified: everything from nose and eyebrow shape to facial hair and expression.

The reason such approach is not good is that the Nazis actually did research on physiognomy and phrenology to show how a person is a bad person (in their opinion). The article in questions says this:

 

Quote

“Physiognomy is now universally, and rightly, rejected as a mix of superstition and racism disguised as science,” the researchers write in the paper’s introduction (physiognomy judged based on facial features, while the more notorious phrenology focused on skull shape). But the pseudoscience was a failure because its practitioners worked backwards. They decided who evinced “criminality” or “sexual deviance” and then decided which features those people had in common.

The point here is not whether the results from the analysis of the photos using this method is correct or not; it only matters whether people believe it to be correct.

[mogandave]

The point here is not whether the results from the analysis of the photos using this method is correct or not; it only matters whether people believe it to be correct.

If someone wants to kill you, does it really matter what the result of the test is?

[onthesoi]

18 hours ago, onthemoon said:

Actually, I do believe it is BS.

 

However, if enough people - or the important people - believe it to be true, it can become really dangerous. Dangerous in the sense that the Nazis would have loved to have a tool like this. 

 

How does bullshit become a dangerous and useful tool for nazis?

[onthemoon]

8 hours ago, mogandave said:

 

If someone wants to kill you, does it really matter what the result of the test is?
 

 

Only if he wants to do bad things to do (doesn't always have to be murder) because the test "proved" it.

[onthemoon]

5 hours ago, onthesoi said:

 

How does bullshit become a dangerous and useful tool for nazis?

People believe the BS, and coupled with right-wing sentiments against gays (also works with Jews or Muslims), they now have "proof" that people who are different form them are bad and need to be exterminated. 

[mogandave]

Only if he wants to do bad things to do (doesn't always have to be murder) because the test "proved" it.

Nothing stops them from claiming the test proved it regardless of the actual results.

[mogandave]

People believe the BS...

Indeed

[onthemoon]

1 hour ago, mogandave said:

Nothing stops them from claiming the test proved it regardless of the actual results.

 

Sure. But now they have "scientific" proof.

[mogandave]

Sure. But now they have "scientific" proof.

Whatever

[onthesoi]

http://uk.businessinsider.com/stanford-professor-thinks-ai-will-soon-be-able-to-detect-your-politics-iq-sexuality-2017-9

[VocalNeal]

^ Artificial Intelligence will soon be able to discover that most peoples self acclaimed intelligence is actually artificial? 

 

Now that is a revelation?

Edited September 13, 2017 by VocalNeal

[onthemoon]

2 hours ago, onthesoi said:

http://uk.businessinsider.com/stanford-professor-thinks-ai-will-soon-be-able-to-detect-your-politics-iq-sexuality-2017-9

LOL! A basic requirement for scientific research is that they publish the memethodology and data, so that other researchers can duplicate and verify ("peer review"). Without this, he cannot be taken seriously as a scientist.

 

And now he can also predict our these:

 

Quote

He predicts that self-learning algorithms with human characteristics will also be able to identify:

• a person's political beliefs
• whether they have high IQs
• whether they are predisposed to criminal behaviour
• whether they have specific personality traits
• and many other private, personal details

Thanks for a good laugh, OP. Is it 1st of April yet?

[onthesoi]

Just now, onthemoon said:

LOL! A basic requirement for scientific research is that they publish the memethodology and data, so that other researchers can duplicate and verify ("peer review"). Without this, he cannot be taken seriously as a scientist.

Exactly!

 

Not to mention the ridiculous idea that the complexities of human sexuality can be condensed down to gay or hetro.

[Oxx]

2 hours ago, onthemoon said:

LOL! A basic requirement for scientific research is that they publish the memethodology and data, so that other researchers can duplicate and verify ("peer review"). Without this, he cannot be taken seriously as a scientist.

 

And now he can also predict our these:

 

Thanks for a good laugh, OP. Is it 1st of April yet?

 

Did you actually bother reading the article? He's not claiming that he can predict those other things, simply predicting what he believes will be possible in the future.  All he's claiming now is that he has been able, through AI, to predict sexuality with pretty good accuracy based upon photographs - far better accuracy than human beings.

 

I presume you also are unaware that the paper has been peer reviewed and accepted for publication in the Journal of Personality and Social Psychology, probably the leading academic journal in psychology.  So of course he can be taken seriously as a scientist.

[mogandave]

 
Did you actually bother reading the article? He's not claiming that he can predict those other things, simply predicting what he believes will be possible in the future.  All he's claiming now is that he has been able, through AI, to predict sexuality with pretty good accuracy based upon photographs - far better accuracy than human beings.
 
I presume you also are unaware that the paper has been peer reviewed and accepted for publication in the Journal of Personality and Social Psychology, probably the leading academic journal in psychology.  So of course he can be taken seriously as a scientist.

10% error is pretty good? I imagine Mose guy men could do better than that.

Also, it uses photos from dating sites. Would those pictures not be particularly gay or straight depending on how you sharpen your pencil?

If men were being executed for being gay, I imagine photos posted online would be significantly different.