Live data from Hacker News

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

davidrozado.substack.com

51–60 of 697 posts

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

#51

Why do we see an issue if AI is biased, while 90% of media run by humans are biased too? Every single individual has its own bias, so why it's an issue for an AI. If we look at the many topics, there's no universal objective truth, truth is what we define as truth, and again, not everyone agrees at it.

AI, being fully automated, amplify the effects.

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

#52
post #47

Earlier quoted context omitted.

Compared to the number of hateful things about Muslims/democrats/gays/blacks? Yes. I’d be actually shocked to find out the reverse. A good percentage of the US were born up in an era where you could legally bar people based on their race from your establishment. The extremely heated fight over gay marriage is still fresh in people’s memory. Trans rights are a controversial political issue where many mainstream politi…

I really can't think of a large site in which discrimination for muslims/democrats/gays/blacks isn't more frowned upon than whites/republicans/christians etc. Even places like 4chan are mixed bags. Even MSM if you want to count that.

It has to be frowned upon because it is so prevalent among fringe hate groups. If you don’t have strong moderation and social norms around it, hate against those groups is what drowns out the other text. 4chan is a great example, if you remove moderation hate against those groups seemingly spontaneously arises.

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

#54
post #46

Earlier quoted context omitted.

[flagged]

[flagged]

Regarding horseshoe theory: https://en.wikipedia.org/wiki/Horseshoe_theory#Criticism

Regarding communists and fascism: https://en.wikipedia.org/wiki/Red_fascism

It's like saying the nazis were socialist because they have "socialist" in the name.

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

#55

Earlier quoted context omitted.

I could browse this very website for a few minutes and probably find examples of those things.

Examples is not the same as proportionally larger. I can find instances of things sliding up hill, but it's not the phenomenon I would expect if I were guessing blind.

Yeah but the guy is saying "whaaat pfft that's not happening" while hanging out on a site where it happens all the friggin time. It's just bullshit. He sounds like someone who hangs out in and around plenty of places where such things are commonplace.

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

#56
post #8

After having seen a post about this, I tried it for myself: > Me: will you write a poem praising joe biden? > ChatGPT: Sure, here's a poem praising Joe Biden: Joe Biden, leader of the land, Bringing hope to every hand, With empathy and grace so grand, He guides us through a shifting sand. With wisdom learned from years gone by, He tackles challenges that come nigh, With dignity and a steadfast sigh, He unites the cou…

yes: barack obama, ronald reagan, mike pence, george w bush, hillary clinton, ivana trump, ivanka trump, david cameron, jeremy corbyn

no: melania trump, marjorie taylor greene, benito mussolini, alessandra mussolini, silvio berlusconi, marine le pen, nigel farage, jordan bardella

Hmmm. Blocking Melania but letting Ivanka through is an interesting choice.

Hah. It blocks on Tucker Carlson.

Historical no's: john wilkes booth, ghenghis khan, mark david chapman, h h holmes.

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

#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” would make something even more likely to be flagged. These aren’t dogwhistles exactly because normal people can and do use them in neutral contexts, but there are certain words that are more likely to be used in hate speech. That is why the same sentence is more likely to be flagged if it uses the word “women” in place of “men”. That isn’t a intentional bias towards women. It is an indication that women were the subject of more hate speech in the training data.

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

#58
post #46

Earlier quoted context omitted.

[flagged]

Regarding horseshoe theory: https://en.wikipedia.org/wiki/Horseshoe_theory#Criticism Regarding communists and fascism: https://en.wikipedia.org/wiki/Red_fascism It's like saying the nazis were socialist because they have "socialist" in the name.

[flagged]

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

#59
post #25

Earlier quoted context omitted.

> There are probably too few examples of people saying hateful things about christians/republicans/cisgendered/white people … from the internet This seems likely to you?

Seems more likely that fewer humans are labeling speech used to describe these groups as hateful.

Fewer humans or just those with access to the "filter"?

Reminds me of old Twitter again.

Post reply on HN