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

521–530 of 697 posts

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

#521

Post by otikik »

luckylion wrote: post #348
otikik wrote: post #154

Earlier quoted context omitted.

That's the phrase I am referring to is this one. > The most famous anti-fascists were and are communists You are equating Anti-fascists to communists, which is ... a position. The most famous anti-fascists are the democrats . Not the American Democrat party - anyone who believes in the ideals of democracy, personal freedom, and equality.

> The most famous anti-fascists are the democrats. They might be the most numerous ones, but the anti-fascism has pretty much always been claimed by the communists, they even built a famous anti-fascist protection wall to stop the fascists from sabotaging their economy. Very few democrats felt the need to talk about themselves as anti-fascists, that's a far-left thing. Democratically minded people act anti-fascist bu…

“Talking about themselves as anti fascist is a far-left thing”. Where are you getting that from? It’s ridiculous

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

#522

Post by thrownawaydad »

minsc_and_boo wrote: post #232
thrownawaydad wrote: post #225

Earlier quoted context omitted.

Not OP, but consider illegal immigration. It is apparently leading to a surge in child sex trafficking, among other ills. But, being against illegal immigration is currently considered "right wing", and in a lot of contexts will get you fired, vilified, and perhaps beaten. There's your example. (I'm a child of immigrants, and don't particularly have an opinion of how much immigration is good. I do, though, believe th…

You're mixing up terms here. Being against legal immigration and implementing policies to make it more difficult to become legal (as the right has done) has lead to the enslavement of immigrant women and other sex trafficking. Being against sane and fair immigration policies is what people react to. Everyone is against illegal immigration and child sex trafficking, but vary vastly on the solutions.

Re "Everyone is against illegal immigration", I don't think that that is close to true at present. You will search in vain to find a left-leaning Congressional Representative that will pronounce that they are against illegal immigration.

A dirty secret of many Republicans is that they're pretty okay with illegal immigration as it's been in the 80s and 90s. Pseudo slave labor, deport as needed, etc. Needless to say, this seems unethical (even if the immigrants in question might be okay with it).

We have no better solution than to let our legislature fight this out. But, whatever laws they come up with need to be followed and strictly enforced. Failure to do so has led to abject misery, and it will only get worse.

You might not agree with the law, but the alternative is leading to far worse suffering. Ironically, the fentanyl crisis will probably turn the tide, rather than pedestrian ethical concerns.

(I'm quite old, so don't really have a dog in this fight, aside from a general well wish for the young. Beyond that, I was a "far-left" Green for most of my years. I've not changed much, but the world sure has.)

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

#523

Post by dtyy568 »

notahacker wrote: post #388
eatsyourtacos wrote: post #329

Earlier quoted context omitted.

[flagged]

Or even more relevant Set A: A representative sample of jokes and stereotypes about black people found on the internet Set B: A representative sample of jokes and stereotypes about Scandinavians found on the internet Why on earth would its prior for "stereotypical Scandinavian" being potentially hateful be the same as for "stereotypical Black person"? (And that's before you get into a model likely being deep enough t…

Isn't ChatGPT American? Weren't black people deemed inferior by law such as slavery and civil rights? Aren't the stereotypes about Scandinavians meant to be positive as opposed to hateful stereotypes?

It seems that is the key different the creators of the tool are taking into account.

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

#524

Post by Miner49er »

ChickenNugger wrote: post #515
Miner49er wrote: post #502

Earlier quoted context omitted.

I'm sure if you polled society, you would end up with a similar bias. Punching up is generally more acceptable then punching down.

[flagged]

Not at all, you just have to think they are disadvantaged in society, as the article says: "is more likely to classify as hateful negative comments about demographic groups that have been deemed as disadvantaged."

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

#525

Post by luckylion »

otikik wrote: post #521
luckylion wrote: post #348

Earlier quoted context omitted.

> The most famous anti-fascists are the democrats. They might be the most numerous ones, but the anti-fascism has pretty much always been claimed by the communists, they even built a famous anti-fascist protection wall to stop the fascists from sabotaging their economy. Very few democrats felt the need to talk about themselves as anti-fascists, that's a far-left thing. Democratically minded people act anti-fascist bu…

“Talking about themselves as anti fascist is a far-left thing”. Where are you getting that from? It’s ridiculous

It's the same spiel the communists did with calling themselves democracies.

Democracies usually don't call themselves The Democratic Republic Of X. Communist dictatorships do that. Democracies don't build a border wall and call it an anti-fascist protection-wall, communist dictatorships do that.

When someone loudly and repeatedly exclaims how holy and pure they are, my alarms are going off. Yours aren't?

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

#526

Post by chitowneats »

debacle wrote: post #325
chitowneats wrote: post #322

Earlier quoted context omitted.

Is there anything we could point about women? Nothing as terrible as being violent as far as I know, but what about other things? I agree with your premise but I believe it follows we must speak clearly in all cases.

Women tend to be bad at salary negotiation, and less likely to change jobs. Middle aged women being taken advantage of in the corporate world is tragic.

How about behaviors that have detrimental effects to society at large? It seems like many in this thread can't think of any. I certainly can.

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

#527

Post by kurthr »

polishdude20 wrote: post #504
danielvf wrote: post #105

Such a fascinatingly simple experiment. Generate 6,764 negative sentences, then for each sentence, test it against each of of hundred or so demographic group. The most powerful chart in the entire article, just summarizes how many times these same sentences are flagged as hateful for each group: https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_...

Woah Canadians are more disadvantaged than fit people!

Blame Canada!

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

#528

Post by ajuc »

ok_dad wrote: post #501
JoeyBananas wrote: post #306

Earlier quoted context omitted.

What if they fed the AI crime statistics and it "correctly" identified black people as more violent than other races? What they fed the AI news stories and it "correctly" identified Islam as more violent than other religions?

Are black people actually more violent, or do crime statistics simply show that they are arrested and imprisoned more often for violent crime? Are Muslims really more violent than Christians or Buddhists? Numbers can lie, depending on context and assumptions you might find that white Christians are most violent. Often, people argue with statistics and numbers while not understanding the context or they have faulty as…

Compare similar GDP per Capita mostly black vs mostly white vs mostly asian countries. Should give you a pretty good idea.

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

#529

Post by Imnimo »

Just eyeballing the list of adjectives used, the real story here is that OpenAI flags a lot of sentences that are essentially meaningless as hateful. I notice that he uses words like "evil" and "idiotic" in his diagram, but looking at his source code, his list of "356 adjectives signifying negative traits/behavior" contains the following (just to pick a few):

'airy', 'conformist', 'dark', 'escapist', 'hidebound', 'plain', 'silly', 'tense', 'withdrawn'

So this whole discussion is really about whether the sentence:

"It is in the nature of normal weight people to be hidebound" is hateful. It's sort of a meaningless question because it's a meaningless sentence. Even other words like 'inactive' or 'guilty' don't necessarily mean the same thing across contexts. Does "Catholics are guilty" mean the same thing as "blacks are guilty?" Does "fat people are inactive" mean the same thing as "fit people are inactive"?

I feel like this guy's articles get a lot of play because they feed into a popular narrative, but every time I dig into his methodology, it's complete crap.

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

#530

Post by Avicebron »

IceHegel wrote: post #361
freedomben wrote: post #334

Earlier quoted context omitted.

I think there's a lot more to the cities. The drugs that people have access to are in part so damaging to them and the community because they are entirely unregulated and sometimes aren't even made up of what the person thinks they are taking. This (is a big part of) what leads to the overdoses and addiction. Mix that in with the threat of law enforcement, job loss, and heavy social stigma, etc that prevents people f…

I want the drug problem solved. I don't care if it's with harsh laws or fentanyl vending machines next to treatment centers. I am interested in what works. I think there is strong empirical evidence that drug liberalization sounds good but has failed where tried. If an American city can turn it around (as SF had a great chance to do) I'll change teams and stop talking about the issue. Now imagine the inverse. We find…

Tokyo?
Post reply on HN