Racist AI at it again

telegraph.co.uk/news/2017/08/24/ai-robots-sexist-racist-experts-warn/?WT.mc_id=tmg_share_fb

Is that because they are correct? it's not the AI that is inherently racist or sexist. It is reality that has inherent differences between sexes and races.
Let's focus more on the science side of AI?

>another retard that doesn't know how modern AI works
If you want to discuss the "science side of AI" then make a thread about linear algebra, or rather just ask in the stupid questions thread.
If you want to discuss the results or other aspects of AI, then go ask Sage for not science.

What does linear algebra have to do with AI?

t.Somebody who has not done LA.

I don't see how an AI is going to interpret anything any other way.

Liberal attitudes of the inner city are dangerous when combined with AI's and their ability to figure out truths.

>AI is literally glorified statistics
>statistics are sexist and racist.

So do people honestly damage themselves by going to the Telegraph to pull bullshit identity politics articles, hoping that their thread gets enough (You)s to take off?

You didn't start this thread with a political agenda at all, no.

That's just it ain't it though.
I fed you data.
''oh you're being political''.

No, REALITY is political.

>When they asked the machine to complete the sentence “Man is to computer programmers as woman is to x”, the machine answered “homemaker”.

Hahahahaha you're retarded for trying to use the opinions of this thing as evidence for your retarded beliefs but I'll be damned if that isn't fucking hilarious

>machine says thing
>look, machine said non PC thing
>this is proof that random boogymen use it for their beliefs
you're reaching a little far there fuccboi

Whoa, it's almost like decades of overt redlining and purposefully confining black people into specific, run-down neighbourhoods to use as de facto ghettoes, and then distributing and encouraging the use of cocaine to undermine black rights (and black power) movements and activists, and THEN taking a viciously hard-line stance against recreation drug use that leads to a high proportion of arrests and thus criminalisation of black adult men - forcing them into contact with harder criminal elements and reducing their competitiveness in the job market, as well as depriving any children of a stable environment to grow up in - might have led to all the black people crammed into one place, poor and messed up.

Who'd have thunk it? Luckily we all know black people are inherently inferior so all of that must have been their fault too.

>AI is perfect
>AI is racist
>By the laws of logical analysis, racism must be perfect
>Heil Hitler

It's also as if people who live in cities, and thus walk past and interact with all manner of people every day of their lives, become more broad minded than farmers living out in the ass end of nowhere surrounded by a handful of people with the same face as them.

There's that, plus the picture posted is bullshit. The crime data is literally cropped as to only show it in the black and Hispanic parts. Quite sad dude, quite sad.

What is racist about saying that there are proportionally more whites eligible for a loan, or with the potential to succeed in academy than blacks ?

Their predictions are skewed by race and gender because the data is skewed by race and gender - BECAUSE the data is collected entirely from within the context of our society. The AI isn't analysing some kind of novel biological data, it's analysing data generated by our society and thus adapting to the fact that surprise surprise, our society is racist and sexist.

It identifies a picture of a man in a kitchen as 'woman in a kitchen' because there are far more pictures and text pieces involving women being in a kitchen than there are men, but that's only because that's the way it is in our particular society. If we lived in a society where there was an equal expectation of men and women to be cooking, there would be an equal number of pictures of men and women in kitchen, and so you'd be holding up the result of this AI analysis as 'truth' that both genders are inherently equally likely to do the cooking, despite the fact that all you've done is hold a mirror up to your own social values.

The loan isn't racist because maybe banks are racist or money is racist? You could just be stating a fact about society. The "potential" thing is racist because that's just like judging people based on skin color. "He's never gonna make cause he's black".
Basically if it's mean, it's racist.

The issue is that rather than actually impartially analysing whether or not a particular person is eligible for a loan or can succeed in academia, the AI has noticed that blacks are offered fewer loans and university places in general and thus negatively weights 'being black' as a criteria for acceptance.

The other issue is that that very difference in proportions of each demographic eligible for loans, academia, parole etc. is that way BECAUSE the previous generation had the same bias against them. It's self-perpetuating, and unless you step in somehow and try to iron out the bias (be it by the ever-contentious affirmative action, or other means) then the next generation will be able to justify its discrimination in exactly the same way. All the while, people have to live with a lower quality of life for their entire existence, just because of the parents they were born to.

kek, i hope this is a troll from /pol/, otherwise this is such a perfect stereotypically ignorant response I have to screencap it.

If each neuron represents a function like y=mx+b, where y is the output and x is the input, you can calculate all of the neurons and the entire training set at once, with a very short amount of code by representing m,x, and b as matricies. It turns out its faster to do it this way than using loops. There is a very good course on coursera called "Neural Networks and Deep Learning" by deeplearning.ai that can be audited for free.

Veeky Forums BTFO

But that's wrong.
The training data was not who was elligible for a loan, but who managed to pay it back. The point was to only allow a loan to people with a high % of payback.
The fact that banks were especially carefull at giving loan to blacks should in reverse positively bias being black since only the top of the barrel was accepted. Yet even whith this strong positive bias, the Neural net still found that being black is correlated with not paying back the loan.

>correlation is not causation
>Be surprised when you use correlation methods to find causes and it doesnt work

Btw /pol/tards have the intelligence of an artificial neural network, black ppl have more relative criminals only because black ppl tend to be poorer (msybe because they started as slaves, duh)

That's not what the data says.
Could it be that's not /pol/tards that are retarded but you instead?

Wrong pic

>Retarded /pol/tard posts proof of correlation as proof that causation exists
',:^)
That data shows that a sub-society of the general American society commits more crime. Black people interact with black people, and black people are usually poor.

This thread is pathetic. It demonstrates a complete lack of knowledge as to how AI works. I'll just say that computers can come to wrong conclusions if they're designed poorly.

Sage this bait.

too bad its saying the opposite

>if the computer comes up with a conslusion I dont like then it must be designed poorly.

>designed poorly
Or given carefully curated data. Either way, factual statements aren't necessarily racist, they're just observations.

>Either way, factual statements aren't necessarily racist, they're just observations.
remove factual. I just woke up and didn't mean for that to be there.

Robots are daft. As in if a kid has a mobility hindering disability like CP they'd figuratively throw them out without regarding the other factors that should be considered or the individual kid themself.

You know the loans given aren't the same most likely so that must be taken into heavy consideration.

they're fucking factual lmao.

stop it.

Thanks a ton user!

You keep finding ad-hoc excuses to justify every evidence thrown at you. You are no better than a creationist or flatard.

Can you think of an experiment that depending on the result will prove to you beyond any reasonable doubt that blacks are inferiors to white ?
If the answer is no, congratulation you are a religious retard.

Experimentation has shown beyond a reasonable doubt that race is a social construct.

Experimentation has shown beyond a reasonable doubt that negroes are on average borderline retarded.

Anti racism is due to empathy a machine has no empathy or any concern for humans at all. The machine sees patterns and works with them any pattern that isnt consistent the machine sees as an error and must be corrected by elimination.

An A.I would deem blacks as feral humans then begin a genocide campaign on intelligent blacks.

I don't know if there are intelligent blacks enough to make a genocide.

You are abusing of correlation to make a point, but have no explanation for that data, by the same logic you can "prove" divorces increases margarine consumption (theres a website with all those silly correlations). It could be evidence for the existence of racism too, you have no proof that your explanation is the only correct one for the observed phenomena.

>still calling inverse convex optimization """AI"""
>caring about it

???

Deep learning is just regression of very large functions with lots of learning parameters. It's like curve fitting, but times a million. It's an easy way to build complicated functions capable of "learning" lots of behaviors through a series of matrix multiply operations.

>geography affects political patterns

no shit

Sad thing is the meatbags will think it is the AI, data, study etc that is wrong and not themselves.

>citation needed

The one true religion that is totally 100% scientific and will save us all. lol.

Liberal ideology runs on the false belief that reality is socially negotiable. That's clearly just wrong.

Good fanfiction my man.

Just as an AI has no empathy toward humans unless somehow programmed in, an AI also doesn't give a shit if there are weird outliers in the data unless programmed to do so.
For an AI to behave like you suggested, you'd have to take a quality check robot from some factory, tell it that dude, instead of us looking for a particular design, how about you just look at the average and do your thing to those that fall outside margins and THEN let it loose. That'd be retarded in a very obvious way.

People get judged by demographics all the time. Young men crash more cars and get higher car insurance. Old people are more likely to die and pay higher life insurance. People with a family history of X illness are more likely to have X.

Applying demographic based prejudices is just pattern recognition, which AI is supposed to do.

>The other issue is that that very difference in proportions of each demographic eligible for loans, academia, parole etc. is that way BECAUSE the previous generation had the same bias against them. It's self-perpetuating, and unless you step in somehow and try to iron out the bias (be it by the ever-contentious affirmative action, or other means) then the next generation will be able to justify its discrimination in exactly the same way. All the while, people have to live with a lower quality of life for their entire existence, just because of the parents they were born to.
But that's not true for poor Europeans, east Asians, and Filipino immigrants. All demographics that do very well within a generation, despite coming from worse situations than western blacks.

>accident
>death
>illness
>potential to succeed

Spot the odd one out. The free world is about giving people opportunity and what not. It might be true the x race is less likely to succeed in y activity but everyone deserves a shot. Plus the reasons behind it may not fit a racist's agenda (I don't think they would). You don't need to insure someone for doing badly in school.

>all black people in all of history across the entire world are former slaves or coke from poverty

>The only reason a computer would give you an answer you aren't comfortable with is because it's poorly designed.

Prove it

How is it the odd one out? You can only maximize your potential to your biological limits. One person with the right proportions will always have better potential to lift than someone who doesn't. On person with brain chemistry more suited to higher learning or spacial understanding or delayed gratification will always have more potential to be a doctor or physicist or simply more financially successful than someone who doesn't.

In no way am I advocating removing someone's "shot". Stop letting your emotions get the better of you.

Conclusions based on correlations aren't observations of real life.

Your race is getting extinct'd and you can't even rationalize your own purposes hahahahahahahaha

Face your replacement, subhuman.

It's the odd one out because no-one needs (it's not absolutely necessary) to know exactly how likely a population is to succeed in a certain activity. The other ones are relevant to an insurance company because you pay them to insure you against these things. I mean unless you're going to use this information to discriminate or something I don't see why you need to pry. Also, there might be instances where it is useful to really define a person's potential at an individual level, say if you were sponsoring them for a sport or competition. It's just about not letting racists run shit because we all know what happens when you do that.

The brain thing is debatable. Having biological potential for intelligence doesn't necessarily translate into success in life (ie. dumb people get rich too).

>It's the odd one out because no-one needs (it's not absolutely necessary) to know exactly how likely a population is to succeed in a certain activity.
Entire business model and location selections depend on it, as well as government programs. It's valuable and practical to have knowelege.

>The brain thing is debatable. Having biological potential for intelligence doesn't necessarily translate into success in life (ie. dumb people get rich too).
Hence why I specifically referenced things like delayed gratification and implied a range of brain chemistry factors. But yes, intelligence, while a fairly reliable factor is predicting someone's success in life, is far from the sole factor. And demographic wide trends don't scale well down to individual levels. BMI is a good example; good for seeing a society wide fat problem, not good for judging individual fitness.

None of this precludes haplotype wide behavioral predisposition differences, though.

Do they really? The government goes: "Hmmm... There's a lot of black people in this neighborhood. We should stop building schools here and start building more basketball courts". Do you see how that could be racist? I'm pretty sure private companies are allowed to do whatever with their data, so they can analyse it a racial level if they want. But that might lead to some racist advertising or something and the last thing they wanna do is look racist I'm sure.

I guess there is real research value to studying differences between races but I have my doubts about large cognitive differences because the brain is so plastic.

Or it could be used to get solutions to better help groups of people who respond differently to different things. Plugging your ears and pretending something isn't real helps nobody.

If that doesn't appeal to you, how about the idea that scientific discovery is worth it just for the sake of scientific discovery. Or maybe the obvious argument that ignoring genetics just lets an actor like China step in and become the world leader, possibly employing much less ethical applications.

The thing is, no difference, multiplied over a lifetime, is small. Making better decisions in life 1% more often has a compounding effect. By the time you're 70, that's a huge difference in your life and impact on others. Multiplied by whole populations, it's the difference between a successful and peaceful society and a hellhole.

Ok I don't disagree with you but society isn't meant to be optimal even if studying races might have merit. I mean solving problems is one thing but creating a "perfect" world is another.

That's why it's not a huge deal if you ease up on race research if it'll maintain race relations.

>When they asked the machine to complete the sentence “Man is to computer programmers as woman is to x”, the machine answered “homemaker”.

This article is hilarious

Birth of a luddite.

>poverty causes crime

whoa ok, being racist to black people is NOT ok but we can call poor people criminals :)

found the american

Poverty does cause crime, but the mistake is thinking it's the sole factor. A poor black man in Japan, China, Africa, North America, South America, and Europe is more likely to be a violent criminal than any Japanese, Chinese, white, mestizo, or Native American person by a healthy margin. We can't run controlled studies, so comparing internationally is as close as we can get.

>Your race is getting extinct'd and you can't even rationalize your own purposes hahahahahahahaha
>Face your replacement, subhuman.

Good job on proving everyone you've got an agenda to propagate in your defamation against these uncomfortable facts, pal.

not it doesn't honeypie

it's crime that causes poverty

>Poverty does cause crime

No. For one, there are such things as rich criminals. For two, there are such things as poor non-criminals.

>poverty causes crime
>therefore poverty is claimed to be then only cause of crime

Stats show rich blacks commit more crime than poor whites.

They evolved in the savagery of africa.
Guess what, whites are more likely to be criminals than japanese.

But you don't see whites denying it or expecting the Japanese to open up their borders.

The fuck? You're literally on a website of weeaboos who dream of living in Japan.

Also show stats from studies.

Go find your own stats. I'm not interested in that debate, i'm interested the problem of AI, and it's relevance to these issues.

What if the AI gathers up all the data, all the info, and comes to the same conclusion as the white nationalists?

That the flooding of blacks does decrease all the things that make civilisation work, and that they were right all along.

Oh wait, by that stage there is hardly any whites left lol.

Chinese AI is also super racist against blacks for some reason. Japanese too.

Because they don't train it to lie.

Is this the IQ denying negro at it again?

Yes you retarded negro, this is all the results of genes. The same way code is the foundation of computers. Genes are the foundation of all biological organisms.

Its almost like blacks are confined to run-down neighborhood because their genes create behaviors that result in run-down behavior. Its almost like blacks are violent because of their genes. Its almost like blacks are more susceptible to becoming criminals because their genes make them that way. Its almost like blacks are biological organisms and humans are not some special creatures whose behavior isn't influenced by genes like 99.9% of every single biological organism in the universe.

Its almost like blacks can have a slightly higher IQ like this faggot, but they still do not behave like whites, because surprise, they aren't equal to us.

yea but ur still reddit spacing
Doesn't matter, AIs will either recognize human society as a reflection of itself or it will try and organize things, fail miserably due to , get a lot of people killed/miserably and eventually reset to status quo.

99.9%?

Sorry mate but it's not 99.9%.
it is actually one of those ONE HUNDRED PERCENT things.
If it's alive, it has some form of genetics.
Even virus's.

normally i say 99% when i assert stuff to.
But in this one example you actually need to say 100%

I think you quoted me (the last one) by accident on there.

>Make no mistake in thinking that I am 99.9% racist. I am 100% pure racism.

Thanks for the clarification. We really care here about namefags like you. Especially one with your credentials.

Racism is what, sinfulness?
Multiculturalism will not save the human race.
Sacrificing whites is no better than the azteks sacrificing virgins to make the sun shine.

Regardless, AI (and ai conclusion denialism) is a danger to every living thing on the planet.

So how about you shove your anti-white rhetoric up your arse, and get on with helping.

>What is an environment.

Occams razor, liberals deny it.

Why are people so scared of IQ and criminal proneness differences between races?
It helps explain why it happens, and it would kick-start a conversation to help us come up with possible solutions to the issue, like targeted genocide, chemical castration or even genetic engineering.


Wouldn't you love to live in a society full of high IQ and civilized people?

The brain is too plastic for the differences measured between races to be strongly roots rooted in genetics rather than in upbringing (you know, nature vs. nuture).

>Why are people so scared of IQ and criminal proneness differences between races?
Not scared. Just aware that race is a social construct, and IQ is a very flawed tool.
>possible solutions to the issue, like targeted genocide, chemical castration or even genetic engineering.
You're not helping your case.
>Wouldn't you love to live in a society full of high IQ and civilized people?
IQ? Don't care. Civilized? Sure, but that has nothing to do with IQ. You're literally saying the people carrying out genocide are the civilized ones lol.

With regard to intellectual capacity I mean.

>IQ is a very flawed tool.

Yet he best predictor of success and health in life, funny how that works.

>You're literally saying the people carrying out genocide are the civilized ones lol.

Genocide is morally sound, Look at Iceland, one of the most civilized countries out there and it genocides all down syndrome babies through abortion. Imagine a world where we aborted any kid sub 130 IQ. It would be the golden age of humanity in a couple generations.


Just like the renaissance surge of greatness after the plagues killed the weak minded.

Animal = wild = low iq = no civilization
African people = wild = low iq = starving without civilization
High iq = civilization

>muh correlations

>Yet he best predictor of success and health in life, funny how that works.
Statistical predictor. Just like BMI is the best statistical predictor of heart disease. Very flawed, but not useless.
>Genocide is morally sound
Totally

>Imagine a world where we aborted any kid sub 130 IQ
>No more arguing with brainlets on Veeky Forums
Damn dude, you've almost convinced me.

What i hate about modern NNs is how inaccessible they are. Not because it's hard to understand, it's just basic linear algebra after all, but because CS people are obsessed with buzzwords and they hide this simplicity under it. By far the most difficult part of learning NNs is translating this buzzword language in papers into mathematics.

>Totally

>Letting mentally challenged people live a life where they don't belong and suffer is more morally sound that not letting them exist atll
Aha, yeah. Now that's some sound morality. Genocide is the way to go, beyond your hurt feelings, there's nothing but benefits for humanity.

>best
Debatable
You only need weight and height and age for BMI.
Affluence might be a better predictor because all you have to do is pull up tax info whereas for IQ you have to take this long test in a controlled environment. And we know affluence is correlated to success in life.

>There are entire races of people that are mentally challenged
I know, it's the blacks, there's no point arguing it with you

I am not a racist, but an elitist. I am fine with all races, as long as they are high IQ and able to live in a modern society without turning where they live in a dangerous ghetto.


There's smart black people, we should isolate their genome as our knowledge of genetic engineer advances and make sure only smart black people (and all others, including whites) are born. Imagine a world with no Niggers, Rednecks, Gypsies, creepy Indians asking you to open bob, no more crazy Muslim goatfuckers or Abbos drinking petrol and sleeping on the roads.
You are silly if you think your petty emotions are above a brighter future for the human race.

>Race is just a social construct
>but a doctor can tell a white man from a black man based on bone structure

I see no logical inconsistencies in those two statements

Nice impersonation of your average social science mayor