Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

nytimes.com

Post reply on HN
Print view View original

141–150 of 173 posts

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#141

Post by jrm4 »

nzeribe wrote: post #139
jpambrun wrote: post #135

Earlier quoted context omitted.

As someone who worked on ml based classification I can totally see a man being misclassified as a monkey on occasion. If you have man and monkey in the output vector, it will happen on from time to time.. If you find a way to achieve 100% accuracy, go claim your Nobel prize..

But that's not what happened, is it? The algorithm hasn't miscategorised men, it has miscategorised Black men. I'd argue the AI has done its job, with a high degree of accuracy. It found what the programmers were looking for when they wrote the code. It was meant to be an insult, and was highly effective. Lack of outrage on HN is as expected; the authors are quite possibly in the comments holding their sides with lau…

Actually I now must strongly depart from your argument of intentionality here, and reiterate something I said before. I was about to respond to the person below, but they're more-or-less right. I don't know how intentional any of the programmers were here, so I'm not going to speculate on that.

But I will absolutely insist that this racist outcome is possible -- even likely -- absent any bad intentions from the programmer(s) who did it.

ALSO: I'm 99% sure of the following:

A black team of programmers working on "distinguishing humans from monkeys" would never let the mistake of "black and not white people identified as monkeys" out the door.

That's the point here. I'm not saying it wouldn't have happened somewhere in the process, I'm saying that you or I (if we don't work on the inside) would NEVER have seen it because it would have been noticed and fixed before then.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#142

Post by jrm4 »

jrm4 wrote: post #134
nzeribe wrote: post #128

Earlier quoted context omitted.

I find it baffling and dispiriting that you, as a Black person, believe this is more likely to be an "error" than an act of deliberate racism. It's getting silly, now. A white racist programmer has done this on purpose because he finds it hilarious. The chances of a computer equating monkey and man, given the correct datasets, are close to nil. There's a blindness - a wilful ignorance - to look away and show good fai…

Yup. Though I would refine it a bit -- it's the "rush to excuse" that's the most dangerous. I don't much worry about the deliberate racist programmer as much as I worry about subconscious deliberate-or-not biases creeping in. This perfectly analogizes to most everywhere else; angry open n-word saying racists are usually powerless losers. More problems come from the larger combination of the rest of the "racism" spect…

And now I'm very curious as to the reasoning behind the downvotes. Again, I'm fine with you disagreeing with me, but I must insist that discussion is more valuable than blind downvotes when you do have the rare occasion of an actual real life black person here giving their opinion.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#143

Post by jrm4 »

jrm4 wrote: post #125
suifbwish wrote: post #121

Earlier quoted context omitted.

I also happen to be black but that actually has very little bearing on what I have to say and am merely mentioning it to prevent anyone from suggesting that my opinion on the matter is related to my skin color. The problem with this is no one actually trained the AI to have a racially bias but rather did not train it on recognizing the difference. Why? Probably because no one knew the computer was going to identify t…

That first sentence, man. Like, I feel bad second guessing people, in some ways it's not my right -- but it so feels like the kind of "respectability" talk I used to say out of what I thought was "the truth," but now realize it was a sort of habitual way to defuse the room. So I will state clearly -- yes, I am aware my opinions are absolutely related to my skin color and that's precisely why you all need to pay speci…

Said it before, saying it again. I merely said "pay special attention," not "always 100% agree." You've got two black folks here with differing opinions ostensibly ready to discuss. I have no problem with being disagreed with, but I'm still good suggesting that mere downvoting isn't helping.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#144

Post by salawat »

dagmx wrote: post #102
Y_Y wrote: post #97

Earlier quoted context omitted.

I've dealt with a lot of badly labelled datasets, but I can't say I ever saw one maliciously labelled by racists. There are other more subtle problems that may contribute. As an amateur photographer I've been told that most cameras are calibrated for light skin. If you aren't careful with your setup you'll get bad definition in shadows and relatively very dark areas. If you have badly calibrated/exposed photos then v…

This is usually a good example of societal bias compounding. It's possible that nobody directly involved in this was actively racist, but we're building on centuries of racism in every facet. So a little bit of bias in each technology and data set involved, not to mention the engineers and their testing, just adds up

>but we're building on centuries of racism in every facet.

Nooo. We're building on techniques that have worked well enough up til now, but people don't bother to dig in a d come to terms with what's actually going on.

Not everything is racist. It's more a tyranny of 'good enough' and convenience.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#145

Post by diskzero »

stathibus wrote: post #22
saltedonion wrote: post #13

Earlier quoted context omitted.

The space for potential edge cases are so large it will be wack mole. Inference tasks isn’t like software testing where the states are well defined.

Another way to put this is: nobody working in this field has control over what they are building, and can't make any promises about how it will behave. It's like launching a rocket ship that you can't test in a physical simulation first. Something is going to go wrong that can't be predicted because we lack the understanding, but this is seen as an acceptable risk.

nobody working in this field has control over what they are building, and can't make any promises about how it will behave. If you are working on machine learning as an employee of Facebook, Google, etc. shouldn't you assume the worst possible outcome from your efforts? I often wonder how people manage this cognitive dissonance.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#146

Post by pure_simplicity »

jrm4 wrote: post #125
suifbwish wrote: post #121

Earlier quoted context omitted.

I also happen to be black but that actually has very little bearing on what I have to say and am merely mentioning it to prevent anyone from suggesting that my opinion on the matter is related to my skin color. The problem with this is no one actually trained the AI to have a racially bias but rather did not train it on recognizing the difference. Why? Probably because no one knew the computer was going to identify t…

That first sentence, man. Like, I feel bad second guessing people, in some ways it's not my right -- but it so feels like the kind of "respectability" talk I used to say out of what I thought was "the truth," but now realize it was a sort of habitual way to defuse the room. So I will state clearly -- yes, I am aware my opinions are absolutely related to my skin color and that's precisely why you all need to pay speci…

> " I am aware my opinions are absolutely related to my skin color and that's precisely why you all need to pay special attention to them. Not because of some sort of silly inherent genetic thing, but because I have experiences that many of you don't."

We used to think of this as racist stereotyping, but in the pendulum of history we have swung back to where in order to not be racist (or, better yet, anti-racist) we have to be actively racist.

The zeitgeist is to view everything through the lens of race / skin color ostensibly in an effort to combat racism, but in reality all it does is create more racism.

The fact that we have multiple black people commenting here with opinions diverging as widely as they do should be proof positive that race or skin color has almost nothing to do with your opinions.

Two people can go through the exact same experience and come out entirely different persons or very similar. Two people can have the same skin color and go through totally different experiences. Two people can have different skin color and go through very similar experiences. Thus when dealing with individuals we should not ascribe to them stereotypes unless we have to make an informed guess, because statistical patterns apply only to groups not individuals.

Guessing my race based on my opinion will be left as an exercise for the unconvinced reader ;)

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#147

Post by pure_simplicity »

jrm4 wrote: post #134
nzeribe wrote: post #128

Earlier quoted context omitted.

I find it baffling and dispiriting that you, as a Black person, believe this is more likely to be an "error" than an act of deliberate racism. It's getting silly, now. A white racist programmer has done this on purpose because he finds it hilarious. The chances of a computer equating monkey and man, given the correct datasets, are close to nil. There's a blindness - a wilful ignorance - to look away and show good fai…

Yup. Though I would refine it a bit -- it's the "rush to excuse" that's the most dangerous. I don't much worry about the deliberate racist programmer as much as I worry about subconscious deliberate-or-not biases creeping in. This perfectly analogizes to most everywhere else; angry open n-word saying racists are usually powerless losers. More problems come from the larger combination of the rest of the "racism" spect…

In this case, it's not the rush to excuse that is dangerous, it's the rush to accuse that is dangerous. Innocent until proven guilty is a fundamental principle of justice. You have not demonstrated beyond reasonable doubt that this is in fact the doing of a racist, so anyone coming up with plausible alternative explanations is only doing so to prevent the discussion from devolving into a witch hunt where the guilt of the accused is simply assumed and our biases and perceptions stand in for evidence. Repeat the witch hunt scenario often enough and eventually you end up with a boy who cried wolf scenario and now you've lost your credibility even in cases where you do have evidence to back up your claims.

When people are careless in making accusations that's incredibly damaging to society. You as a black person should know how divisive and damaging false allegations can be, since that was a tool racists used to instill racial hostility in society. Let's not repeat those mistakes. I'm not saying racism doesn't exist or that we shouldn't do anything about it, but knowing in your heart of hearts that this must have been an act of deliberate racism is simply not enough. We need a smoking gun, or else it's best to withhold judgment.

The fact that no one thinks the mislabeling by the AI is acceptable should be enough. That shows you that no one would defend a person who deliberately designed this. The 'rush to excuse' is not because society doesn't want to condemn racism, but because most of society still cares about fairness and not making false accusations.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#148

Post by jrm4 »

pure_simplicity wrote: post #146
jrm4 wrote: post #125

Earlier quoted context omitted.

That first sentence, man. Like, I feel bad second guessing people, in some ways it's not my right -- but it so feels like the kind of "respectability" talk I used to say out of what I thought was "the truth," but now realize it was a sort of habitual way to defuse the room. So I will state clearly -- yes, I am aware my opinions are absolutely related to my skin color and that's precisely why you all need to pay speci…

> " I am aware my opinions are absolutely related to my skin color and that's precisely why you all need to pay special attention to them. Not because of some sort of silly inherent genetic thing, but because I have experiences that many of you don't." We used to think of this as racist stereotyping, but in the pendulum of history we have swung back to where in order to not be racist (or, better yet, anti-racist) we…

A little exhausting to, as always, have to thread the needle and calm things down, but let's try again.

My insistence on what I said isn't about anything permanent or immutable. This is about what I hope to be a temporary state of affairs, namely that -- it is at least plausible that the lack of black folks in tech today is leading to unintended negative consequences, and perhaps we should talk about that directly without having to worry about the fragile torpedoing the conversation by confusing this hopefully temporary state with some bigger overarching "statement about people."

And maybe I didn't quite say this in the exact way that won't set people off the first time, but you know (or should know) that we have to do this A LOT. I was hoping to be able to skip it, but hey, guess not for now.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#149

Post by jrm4 »

ant6n wrote: post #136
jrm4 wrote: post #100

Black person here. So, the lack of outrage here (I mean in hacker news) is unsurprising, and would of course be greater if there were more black people "in tech" and by extension here. And even as I write this, I'm not trying to judge -- just more open the door to understanding, like, y'all understand this fact, right? I find myself searching for metaphorical equivalents. I'm thinking something like "what if it label…

I find a lot of the discussion here baffling and tone deaf. There are a lot of comments saying "but humans are primates". It's a dismissal of systemic racism issues coming from a privileged position that's akin to responding the black lives matter movement with "all lives matter". Some of the commenters may exactly think that the algorithm is objective, forgetting that it's built by ppl, for ppl, tested by ppl. If th…

Sadly but unsurprisingly, I'm seeing it mostly isn't. But we soldier on.

Re: Facebook Apologizes After A.I. Puts “Primates” Label on Video of Black Men

#150

Post by pc86 »

skinkestek wrote: post #137
pc86 wrote: post #126

Earlier quoted context omitted.

Fined for what, exactly? They deployed a flawed AI model. The flaw was pointed out, and they removed the model and apologized. I'm genuinely curious what else should have happened.

What should have happened is they did basic testing up front? And since they fixed it right away small fines only, like nuisance small. And small follow up.

You didn't answer the question. Fine for what? Are you suggesting broken AI models should incur criminal charges? What exactly are they being fined for?
Post reply on HN