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

Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

breakermag.com

Post reply on HN
Print view View original

321–330 of 388 posts

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#321

Post by imh »

ForrestN wrote: post #298
tomp wrote: post #146

Earlier quoted context omitted.

More to the point, “procedural fairness” or “equality of opportunity” (if I understand it correctly) is directly opposed to “representative fairness” or “equality of outcome”. In many situations, prioritising one will compromise the other. As an example, Harvard admission of Asian students. Perspectives differ on which one is more important. Personally, I’m absolutely for the former, but it increasingly seems I’m a p…

I think the most useful way to think about this is that unequal outcomes almost always lead us to unequal opportunity upstream. If, for example, you made admission to elite colleges(which is a whole other can of worms) ignore ethnic backgrounds, you'd get a result where far fewer applicants of certain ethnicities would be accepted. In an arbitrary sense, this could represent "equal opportunity" if you believed that t…

It's great to account for what structurally led to the moment of a decision. What's tricky is that it's hard to get the right level of granularity. If you don't have data on how historically disadvantaged a given person's path was, you're going to get false positives and false negatives, and you can't make the system fair across both.

Changing the discussion to be structurally fair, not just transactionally, you still have the same impossibility issues arising if you don't have an oracle model with omniscient features. It's the same issues, but with a larger scope of features and outcomes.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#322

Post by dang »

gdy wrote: post #301
dang wrote: post #254

Earlier quoted context omitted.

We've banned this account for trolling. That's what flamebait on inflammatory topics is, intentionally or not.

He has a point, though.

It seems like a shallow and obvious one to me, but if I'm mistaken and there's more there there, it deserves to be expressed substantively and respectfully. Otherwise the only effect is to rile up agreers with positive charge and disagreers with negative charge—just what we don't need.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#323

Post by DuskStar »

fzeroracer wrote: post #309
tomp wrote: post #305

Earlier quoted context omitted.

> which by your argument would be racist. You seem to have misunderstood my argument; maybe I wasn’t really clear in that part. I think a policy that helps lots of poor black people, some poor white people, and doesn’t help Obama’s daughters (because they don’t need help), is a good, fair, and most importantly effective policy. A policy that doesn’t help poor white people is racist. A policy that helps Obama’s daught…

A policy that affects one group of people more than others by racial lines sounds like a racist policy, don't you think? Which was my entire point. You're in effect, arguing for a policy that is racist by your own prior argument. We have affirmative action policies and the like because we admit that some groups of people are disadvantaged due to a multitude of circumstances and we attempt to do things in such a way a…

I think you're conflating the two different definitions of racism here. In one, something is racist if it has different impacts on people of different races - not due to their race, but due to some correlated variable. In another definition of racism, something is racist if it uses race as a determining factor in behavior or treatment.

Providing aid to low-income families is racist by the first definition, but not by the second. But (damn near) everything is racist by the first definition.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#324

Post by anon4justice »

kauffj wrote: post #11

There is no design of AI that can be simultaneously utilitarian, procedurally fair, and representatively fair. I'm going to repeat this again, since many people struggle with this. It is _literally impossible_ to achieve the best utilitarian outcome, the most procedurally fair outcome, and a representatively fair outcome. Anyone designing algorithms will have to make tradeoffs along this frontier. For a far better ar…

The author of these slides is a known racist. He was banned from HN for posting racist comments: https://news.ycombinator.com/item?id=14115498

His slides reveal his biases. Slide 41 cites a graph taken from a white supremacist twitter account (if it's not a Russian troll).

It's possible that AOC made a couple of minor computer science mistakes. But it's very clear that she's morally right and all the dissembling and racism apologia here is pretty sickening.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#325

Post by adlkjnndnnd »

hannasanarion wrote: post #237
adlkjnndnnd wrote: post #234

Earlier quoted context omitted.

What makes you think that? No, of course not. I'm just saying having read that gorilla story doesn't imply any degree of technical understanding. Also, I have to say I considered it hyperbole to be outraged about the Gorilla. It seems pretty obvious that it was just a mistake with the data, not intentional. It is a good warning for things to watch out for, but there wasn't really anything racist about it. Iirc there…

>I'm just saying having read that gorilla story doesn't imply any degree of technical understanding. Is great technical understanding required before one can evaluate whether a program that labels black people as gorillas is functioning appropriately? >Iirc there are even physical reasons why it is more difficult to identify black faces than white faces. Is that then racist, if an algorithm struggles more with identi…

"So you are saying" - are you saying you are really a parody/troll account, referring to that infamous Jordan Peterson interview?

No, I am not "saying".

My comment was in response to praise of AOCs alleged technical understanding, not of her ability to judge the gorilla algorithm.

And you don't seem to understand what algorithms do. A simple algorithm could count pixels in an image. If most pixels are white, it could say "human", if most pixels are "black", it could say "gorilla". It would be a verify bad classifier, that would only work in a number of cases. For that algorithm, you could say a black person would be more similar to a gorilla. But nobody would "be saying" black people are similar to gorillas, just that the algorithm would be more likely to classify them as such.

Are you saying people would use the "authority" of such an algorithm to claim black people are gorillas?

"If you're trotting out that grand old "black people all look the same" thing"

I didn't - stop imagining so many things. I am not a photographer. I think there were issues with the lighting and contrast. Physical issues. Other commentator claims it is just because film equipment was calibrated that way.

Even then I would dispute the "racist" label. There are many different looking people on the planet. Just because you can not account for all of them, it is not racist.

I am inclined to call your attitude racist, because you assume everybody is surrounded by the same mix of people (like in the US), and maliciously chooses to ignore certain types. That overlooks the reality of people who are not surrounded by an even mix of people of all types.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#326

Post by adlkjnndnnd »

billy_beef wrote: post #267
adlkjnndnnd wrote: post #234

Earlier quoted context omitted.

What makes you think that? No, of course not. I'm just saying having read that gorilla story doesn't imply any degree of technical understanding. Also, I have to say I considered it hyperbole to be outraged about the Gorilla. It seems pretty obvious that it was just a mistake with the data, not intentional. It is a good warning for things to watch out for, but there wasn't really anything racist about it. Iirc there…

> Iirc there are even physical reasons why it is more difficult to identify black faces than white faces. Is that then racist, if an algorithm struggles more with identifying black faces? Unless there is some physics based reason it is orders of magnitude more difficult to do feature detection against black skin, then yes something racist is going on. The tools/cameras that we've designed have historically been meter…

Then I would say racial bias is not the same as racism. You can not really blame the people who developed those tools for not being surrounded by people of all races, or for the use case of, say, white people. It's not a malicious intent, whereas racism implies malice.

It seems to me many US citizens think the whole world is like the US. It isn't. There are many countries where there are not as many "races" living together as in the US.

I think it's an important point. For example, if I were a writer and I would write a novel, I would perhaps only feature white people in it. But not because I have anything against black people, but simply because I don't really know any black people and would therefore be hard pressed to write about them. Doing that wouldn't make me racist - I would simply write a novel from "my" world. It would in fact probably be impossible to write a novel that accounts for all possible human experiences.

Edit: would it even have been possible to calibrate chemical film to work equally well for black and white skin? What if it would have been necessary to have different film rolls for white and black skin, for optimal result. Would those have been racist film rolls? Or what about makeup for different skin tones? Is that racist?

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#327

Post by fiter »

billy_beef wrote: post #289
fiter wrote: post #272

Earlier quoted context omitted.

The FICO example from another thread shows a different outcome based on race without any data in the computer system about race. If the computer system doesn't have access to race, like in the FICO example, then can it be discriminating based on race (racist)?

Yes, if the data contains proxies for race and/or unexamined racialized bias, then yes it can. Police presence has often been kept higher in zipcodes with mostly non-white residents. This results in higher arrest rates in an area because of increased eyes, not necessarily because of increased rates of law-breaking. If an algorithm were to be designed to determine where to allocate police officers using this data, it…

My apologies for not linking to the other thread, but as far as I can tell, there are no proxies for race used to calculate FICO scores.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#328

Post by Domenic_S »

rtkwe wrote: post #246
50656E6973 wrote: post #173

Earlier quoted context omitted.

Equality of outcome seems like a mechanistic concept of the industrial age of standardized mass production. Equality of opportunity seems much more organic/humanistic because it allows for the natural variation in people's values and personality

The problem is equality of opportunity is it's invariably applied in a way that ignores some very racist and uneven starting conditions and the effects of the past. Consider as a small example how schools are funded in the US, which is largely by local property taxes, so poorer places have poorer schools. These areas are heavily skewed towards minority populations because of some extremely racist issues in our past l…

> Consider as a small example how schools are funded in the US, which is largely by local property taxes, so poorer places have poorer schools.

Varies from state to state. In CA, only about 25% of K-12 funding comes from property taxes: https://ed100.org/lessons/whopays

I know of several school districts nearby that contain nothing but relatively new $1mm+ houses and abysmal school ratings. I would believe that there's some kind of complex nonlinear correlation between property values and quality of school, but it is very clearly not an easy "more expensive houses == better schools" calculation.

> perhaps just maybe it comes from decades of lack of opportunity and systemic abuse?

A fine point, and certainly not limited to any one race. I'm all in favor of helping those with fewer opportunities growing up, and I don't really care what race they are. If they happen to be mostly $RACE, it doesn't matter to me.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#329

Post by fiter »

ForrestN wrote: post #312
tomp wrote: post #304

Earlier quoted context omitted.

Can you please explain how Obama’s daughters are disadvantaged based on ancestral violence? The point is, racist discrimination, even if “positive”, produces unfair outcomes. Obama’s daughter don’t need even more advantages. If you want to help kids from disadvantaged backgrounds (and I think that’s a worthy cause, even if you ignore “fairness”, simply because it increases the odds of finding the next Ramanujan and i…

Oh wow! Yes, that's very easy! Let me count the ways. First of all, Obama's daughters are much more likely to be murdered, abused or falsely accused by police than they would be if they were white. Even iconically famous black people are sometimes mistreated by police officers who don't recognize them. I'd say being at a high risk of murder counts as a disadvantage! I don't even need to draw a line back to ancestral…

What if Obama's daughters don't find being more likely to be murdered to be a worthwhile exchange for whatever advantages you're offering them? It doesn't seem like a fair negotiation to not include them.

Re: Alexandria Ocasio-Cortez Is Absolutely Right About Racist Algorithms

#330

Post by Domenic_S »

sam0x17 wrote: post #287
kauffj wrote: post #27

Earlier quoted context omitted.

Yes, it is. But if you're designing an algorithm that assesses credit risk and you do this, you will lose money compared to a company that does not do this.

Frankly, private industry has been lucky up until this point that anything other than your financial history has been allowed to factor into your credit score or credit worthiness __at all__. I wouldn't be surprised if it gets limited to solely financial factors by regulation within the next 10-20 years. In terms of regulations, I think by regulation they shouldn't be allowed to give the algorithms anything but loan…

I'm not sure what you mean; are you saying an otherwise-creditworthy person could get denied a loan based on other factors?
Post reply on HN