Robots Are Racists

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
Just so you know, robots are racist. Either that, or they have a fetish for white chicks. Either way, they are still going to KILL US ALL!

Beauty.ai, an initiative by the Russia and Hong Kong-based Youth Laboratories and supported by Microsoft and Nvidia, ran a beauty contest with 600,000 entrants, who sent in selfies from around the world—India, China, all over Africa, and the US. They let a set of three algorithms judge them based on their face’s symmetry, their wrinkles, and how young or old they looked for their age. The algorithms did not evaluate skin color. The results, released in August, were shocking: Out of the 44 people that the algorithms judged to be the most “attractive,” all of the finalists were white except for six who were Asian. Only one finalist had visibly dark skin.
 
"... an initiative by the RUSSIA and HONG KONG-based Youth Laboratories ..."

" ... all the finalists were WHITE except for six who were ASIAN ..."

Russia -to- white.

Hong Kong -to- asian.
 
Blame the programmers, they wrote the algorithms.

Besides, it's not racist to prefer/like what you are used to or have been exposed to most your life.
Just because I don't like I don't like certain types of music that certain minorities seem to prefer, isn't racist, it just means I have a different taste in music.
(There is music produced by people who look like me I don't like too, so does that make me self-hating?)
 
What, you don't like 80-90s rap? You're a nutzo. Or just waaaaaay too old.
 
This will end well...
St%C3%B6wer_Titanic.jpg
 
All those times I said "Kill all humans," I'd always whisper "except one".
 
So A.I. is as racist as the rest of the world. Hmm what a coincidence!
Maybe the A.I. is what is running the international sex trade. They seem to have the EXACT same values.
 
I went and actually looked at the pictures of the winners. Looking at the criteria they are working from, I can see how they were ranking well on those, but you can score well on those and still be ugly. It's not a beauty contest.

As far as the racist part of the algorithm goes, it's right there on the webpage:

"PIMPL, analysing the amount of pimples and pigmentation"

/facepalm
 
The word racist itself is turning into a great qualifier word that will help me quickly and easily spot people that are fucking stupid. Seems more and more the word "racism" is being applied to any situation where one group does not like the outcome.

Kinda like with the boy who cried wold, the more you use it... the less meaning it has. Topping all that off, me, you, the programmers of that robot are all allowed to have personal preferences when it comes to beauty. That doesn't make me racist... it means I like what I like.

I happen to find midget women from the congo very very unattaractive... queue up some loudmouth protester saying I'm racist. *eyeroll*
 
Nothing would have been mentioned if the results were flipped.
There is only ONE group who are told they are racist.
Lets watch some NFL games to see if we can all figure out who everyone is condemning world wide. Hint, it might be during a national anthem.
If your white. Ya can't be right!
 
HP had this embarrassing problem when they launched their face tracking webcams on all their products that wouldn't recognize blacks as people. It was tested by asians and whites, and dark skin lacked sufficient contrast to recognize facial features since it couldn't see the shadows and darker eye areas in relation to cheeks and what not, which wasn't noticed until after launch.

However, if you read this article, you can see its clickbait. Almost all the contestants were white, and only 1% were of African heritage.

If almost everyone in your contest is white, then obviously you have a high probability of only picking whites as your winners.

DERP... ;)

If only 1% of contest entrants were white, and the winners were all white, then you'd have a reason to second guess the algorithm.
 
HP had this embarrassing problem when they launched their face tracking webcams on all their products that wouldn't recognize blacks as people. It was tested by asians and whites, and dark skin lacked sufficient contrast to recognize facial features since it couldn't see the shadows and darker eye areas in relation to cheeks and what not, which wasn't noticed until after launch.

However, if you read this article, you can see its clickbait. Almost all the contestants were white, and only 1% were of African heritage.

If almost everyone in your contest is white, then obviously you have a high probability of only picking whites as your winners.

DERP... ;)

If only 1% of contest entrants were white, and the winners were all white, then you'd have a reason to second guess the algorithm.

Exactly.

If 1% of the entrants are of a specific color, even without any bias (or criteria for that matter, it could be straight random) you would have 1% of the winners being that color, and 1% of 44 winners isn't even half a person.

However, that same rule should have meant that 75% of winners were white, when it ended up being more like 86%. It was the Asians who really got the short end, with a low entrant/win ratio
 
Back
Top