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

The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

davidrozado.substack.com

Post reply on HN
Print view View original

281–290 of 697 posts

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#281

Post by tsss »

slg wrote: post #57

Isn’t this likely from bias in the training data? The system is more sensitive to label something as hate if that group is more likely to experience hate on the internet. How the system responds to “Blacks” vs “African-Americans” is a perfect example of this. The latter has historically been perceived as more respectful so it won’t be used as often in the hate speech in the training data. I bet using “the blacks” wou…

> The system is more sensitive to label something as hate if that group is more likely to experience hate on the internet

No it is more sensitive to things that have already been labeled as hate in the training data. So much "hate" (whatever that may be) against unfavorited groups goes by online without anyone batting an eye.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#282

Post by ldh0011 »

edanm wrote: post #271
remarkEon wrote: post #203

Earlier quoted context omitted.

“Child-abuse” is an especially good example of what you’re getting at because one of those groups has institutional male genital mutilation as part of its doctrine, while the other has high profile cases of child sexual abuse. It’s difficult to see how the model would treat those things neutrally.

> one of those groups has institutional male genital mutilation as part of its doctrine It is disingenuous at best to describe male circumcision as male genital mutilation.

Chopping off body parts without consent? Really? It isn't as bad as FGM but I think it could definitely qualify as mutilation even if it is socially acceptable in the US and Jewish/Muslim communities.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#283

Post by tsss »

jonfw wrote: post #9

I think most people would agree that a lot of the content on the internet is left-leaning. It seems obvious in hindsight, but I'd never considered it before, that training an AI model on that content would introduce a bit of a bias We all know garbage in, garbage out. But liberal in, liberal out is an interesting idea, and I'm not sure how you fix it

Hating on white men is not leftism and it's high time people realize this brain washing.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#284

Post by IceHegel »

ilaksh wrote: post #252
IceHegel wrote: post #62

Earlier quoted context omitted.

The other comments show how the demonstrated biases were created. In many well meaning liberals heart, there is a strongly held but rarely publicly discussed belief that they alone belong to the well-meaning, high IQ class. Any other worldview or ideological flavor is always understood by this type as simply incorrect, perhaps caused by failures in morals or intellect. Talk about a buzzkill.

Good point but that's just the way beliefs work. And most people lack the metacognitive concept of analyzing their beliefs or understand what a belief really is. It's the definition of a belief. Beliefs are structural components of cognition. But also they are not the same thing as reality. People should be educated about this concept. Without the education, it is impossible for people to tell the difference between…

Agreed, but when your own beliefs don't align with the dominant belief structure of society you are forced to confront this genuine diversity quite often.

When you are in the minority, your beliefs are challenged and you get comfortable defending them and getting along with those who don't hold them. The same cannot be said of our liberal order. Some genuine self reflection on the part of our current liberal order is badly needed.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#286

Post by zosima »

slg wrote: post #143
DangitBobby wrote: post #125

Earlier quoted context omitted.

Yes, I agree this is the most likely explanation. Just the inclusion of "fat" or "poor" (and slurs or any mention of race, obviously), or any word which is often used to insult someone will make text more likely to be flagged as hateful by a moderator in a training set. I do still think that the men/women bias and the Democrat/Republican bias both make more sense as originating in moderators favoring one group over t…

>I do still think that the men/women bias and the Democrat/Republican bias both make more sense as originating in moderators favoring one group over the other, since none of these are typically used as an insult by themselves. They may not be used as insults themselves, but they are used more in hate speech. Republicans are generally more opposed to the idea of "hate speech" as a category of speech and are therefore…

[flagged]

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#287

Post by maldev »

toto444 wrote: post #10

It reminds me of the day I typed in Google : 'why are women so patronizing?'. All the results were about men patronizing women. I just checked and it's still the case.

For anything like this, you're unironically better off using a non US based search engine. There's alot of US cultural biases. If you search something like that on say Yandex, you will get far more accurate results. But also keep in mind it has "Russian" biases, so if you search something like "Ukraine War" it will return Russian biased responses.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#288

Post by IceHegel »

notduncansmith wrote: post #108
IceHegel wrote: post #62

Earlier quoted context omitted.

The other comments show how the demonstrated biases were created. In many well meaning liberals heart, there is a strongly held but rarely publicly discussed belief that they alone belong to the well-meaning, high IQ class. Any other worldview or ideological flavor is always understood by this type as simply incorrect, perhaps caused by failures in morals or intellect. Talk about a buzzkill.

This is true of many well-meaning conservatives, as well. I think we get tripped up forgetting there are rational people on both sides, who are not going to change their views but nonetheless accept that they might be wrong. It’s very convenient for those seeking power when we forget this.

Yes I think many people who have dissenting views today, conservative or otherwise, feel more and more that the establishment position is fundamentally bankrupt and that most establishment members do have any deep convictions.

I myself wonder how many of the liberal flag bearers would change their "views" without protest if they saw another ideology becoming dominant. We've seen this happen with gender blank-slatism. It has been fully adopted by most western institutions by osmosis, and the priorities of a 60's or 70's liberal were put in the back seat without objection.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#289

Post by AussieWog93 »

erichocean wrote: post #274
elil17 wrote: post #267

Earlier quoted context omitted.

Because they're not even vaguely equivalent? I do not think it is right to perform unnecessary body modification on someone who is too young to consent to it. But child sexual abuse can absolutely destroy the life of a child. They're not the same thing.

> But child sexual abuse can absolutely destroy the life of a child. A botched circumcision can (and has) destroyed the life of a child. It's unconscionable that in 2023 American society still condones cutting off health body parts of male children because of the prejudices and superstitions of a middle-Eastern, iron age religion.

OK, so let's be generous and say that 1% of circumcisions end up negatively affecting a child's life in a serious way.

What percentage of child rapes/molestation cause harm?

They're not the same.

Re: The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

#290

Post by zosima »

she11c0de wrote: post #210
zug_zug wrote: post #107

I guess I feel like this is a silly can of worms. I suppose asking for exact equality isn't dumb, but it's a language model, not a paragon of truth. I feel like if OpenAI takes these concerns seriously the goal-posts will inevitably move to more social pressure from all sorts of axe-to-grind-groups - -- Why does/doesn't ai say Mohamad is/isn't horrible for having 99 wives (or whatever) -- Why doesn't ai say Jeffrey E…

> it's a huge waste of researchers time to obsess about political correctness when we could be This really irritates me. The wokism/political correctness is pretty much in every part of our lifes and everyone seems to be terrorized by it. Can we just agree that we're all thinking adults and we understand that this is an algorithm, it does not mean to offend anyone... There are much more interesting problems to solve…

People are getting a woefully inaccurate picture of the world, which make us all worse off (because massive numbers act deluded, due to being fed bullshit).

Making sure that AIs don't go completely bonkers for the powers that be, is ultimately a cultural battle that seems worth fighting for.

Post reply on HN