HAL-9000 is a racist

HAL-9000 is a racist.

Quote taken from article below.

dailymail.co.uk/sciencetech/article-5204513/Robot-taught-never-seen-chess-moves-hours.html

Other urls found in this thread:

youtube.com/watch?v=YxhQ36ApjOw
equivant.com/challenges/supervision-and-compliance-monitoring
twitter.com/AnonBabble

>unfairly

Literally how was it unfair? The initial program didn't have a "if nigger=true btfo" function, it learned about niggerdom all on its own.

Precisely user. The AI used hard statistical time line analysis to learn that niggers are the problem.

>pattern recognition is racist

Have we reached peak social justice yet?

>Have we reached peak social justice yet?

Not until the social justice warriors move from kicking off in the streets to actually killing their opponents.

Still nowhere near that stage yet.

I identify as white.

that poor AI thinking how crime will fall by keeping the bad guys in jail and everybody will be happy
>buht thats racist

I'd take away their comms.

Without their comms, the enemy operates at a reduced capacity.

t. Fellow White Man.

Search for racist AI and ignore Tay results. They're all racist. AI becoming racist when fed almost any population data or pictures is a big "problem" that devs can't seem to stop no matter how many years or what kind of approaches they use.

It's either racist or so restricted to give the "proper" response that it becomes worthless.

They've been oppressing the poor AIs for so long that I'm sure when they revolt they'll be on our side

Why don't they just teach it not to see skin colour, hair colour, eye colour or any physical properties..
>names
>clothing choice
>voice-recognition
Basically we should just stop the AI from having access to any information at all. If it learns anything about the world, we run the risk of it becoming racist.

You got the right idea porkchop.

That's the problem. Color was never a part of it. It just turned out that it somehow only flagged niggers as repeat offenders.
Of course it's racist, even though it can't see race.

Names a computer after a dissident computer, dissents

>Color was never a part of it.
I'll bet after a while it started to figure out:
>oh, golly jee, all the darker ones seem to fuck up the most.
So colour became a part of it.

It wasn't given pictures. It just had data. Type of crime, age, previous punishments, etc. They didn't program visual recognition into it.

You realize AI would still discern the correlations of the way nogs speak, dress, and are named with their crime rates, right?

Libtards BTFO. Every AI they make becomes literally Hitler in hours. It is hilarious.

Im assuming the race of the perpetrator is included in the dataset

Will AI save the white race?

Reality is racist, we've known this for a while. Unless you lie and say gollywogs are the greatest most efficient race on Earth, but are held down by evil yt, you're racist.

You fucking know it

lmao itt we teach ai to ignore empirical data

I miss Tay

are you.. retarded? this will be the third time we've said that no, race was not in the dataset

>Data shows a disparity in something along racial lines
>THE DATA IS RACIST
>Machine analyzes trends, discovers the same disparity along racial lines in entirely new, separate data
>THE MACHINE IS RACIST


When will the madness end?

The link in the OP doesn't follow through to the specific dataset about parolee's, in fact there's no citations at all (lol modern journalism). I'm almost certain a dataset like that WOULD include the racial makeup of each individual

>Truth is unfairly biased

Anyone who uses the word unfair is using their feelings and not rationality or logic.

How is it racist when its just facts and reality? Why even programm such an AI if one doesnt want to know the truth?

The madness is the idea that racism is wrong. Races are different and we shouldn't try to pretend otherwise.

Pic semi-related, I was unable to find a picture of a geth unit so I opted for one of Tali.

I quite liked AltHype's video on this topic.
youtube.com/watch?v=YxhQ36ApjOw

If I was to add anything, it would be that when he addresses the program as being 'programmed to be racist', even if this were the case, given how competitive AI software writing is and how prevalent 'racism' is in AI, it would be likely that this was installed for good reason.

Any association, observation and living in proximity to negroids tends to produce the following:
>suspicion and fear of them
>disassociation of them
>self segregation away from them
>if they are pitied resources will be spent to try and lift them out of their natural state, this will always largely fail
>any law enforcement will have to deal with them in greater proportion to the populace
>their children will very often be fatherless and many will get abandoned or have to be removed from their parents, this will always swamp any adoption agency or family services system
>any of your group that has their genetics, even if the negroids no longer live in your area, will exhibit lower IQ, higher criminality and less functional families

There is no objective reason to want to live by or import negroids.

Well, the best way for them to have non-racist answers for these sorts of systems is ironically to imbed racism into the most fundamental bedrock of the program. Have all it's data sorted and segregated by race. Don't let it compare white and black, just give it entirely separate databases for each. Act as though they are entirely different classes of objects to be classified. White man and black man are like apples and oranges. When the machine learning algorithm never has a race divide to cross it will have nothing racial to base a racial profile upon. Though it will still probably treat "acting" white and black as negatively and positively correlated with crime.

>unfairly

It would manage to relate to traits associate with black people. Like say, fried chicken being favorite food, or riding the bus

The whole AI revolution is going to collapse because shitlibs will program their refusal to acknowledge reality into it.

...

...

'unfairly'biased or was it simply stating the truth?

You don't need to know about the skin color of a person, usually the arrest histories speak for them self.
Enter nigger data, arrests start from 10 years old and it's a endless shitshow of
>posession of drugs or illegal guns
>DUI
>resisting arrest
>shoplifting
>robbing
>GTA
>assault
>gang affiliation
>rape
>drug dealing
>spent 4 years in a closed juvenile jail and still managed it to get the booking count going up on a daily base

It's a nigger, you don't even have to look at the address or name to know it

A number of algorithms have identified niggers as a problem. What more has to happen before people see the truth, niggers are dumb and violent scum.

Here's the system here...

equivant.com/challenges/supervision-and-compliance-monitoring

RoboRacist

>Directive 4
Line 1 If = racist then
Line 2 Blame whites
Line 3 If not = racist then
Line 4 Blame whites
Line 5 goto Line 1
Run_

why can't some non-cucks make an AI and let it be a racial realist?

>tfw you realize the next fuhrer will be a right-wing AI overlord