Which A.I. program did you use to write this? I think there are some bugs in the system?
It took just one weekend for Meta's new AI Chatbot to become racist
At least it's not sentient.
Which A.I. program did you use to write this? I think there are some bugs in the system?
You have a gift for snark and snot. But Substance?
How and why do you think that has been “done.”
Critique of wokeism from the left…
I don’t disagree with the general premise, but the article isn’t great. Remember when Obama called out wokeness in 2016? How does that fit?
It was a political ploy, designed to boost the chances of his cronies and their capitalist donors, and without any real impact.
It reads woke.
Really? he sounds more like a communist
There’s a thread on this!
He doesn’t sound liberal. And probably doesn’t want to.
I wonder how much of the ChatGPT wokeness is hardcoded, and how much is just its learning set. I mean, if you imagine googling the same set of questions pre-2021, I feel like the results you’d get for “how could jewish/black people improve themselves?” would be … pretty different in tone.
I wonder how much of the ChatGPT wokeness is hardcoded, and how much is just its learning set
I’d say it leans hardcoded, look at these 3 other examples
At least it's not sentient.
The Verge is about technology and how it makes us feel. Founded in 2011, we offer our audience everything from breaking news to reviews to award-winning features and investigations, on our site, in video, and in podcasts.
Researchers at the Allen Institute for AI created Ask Delphi to make ethical judgments — but it turned out to be awfully bigoted and racist instead.
I guess what I’m imagining is a filter that says “If the generated response is racist, show this instead”. But my point is that the suggestions it spits out for white people aren’t themselves racist – we should all practice empathy, etc. And there’s loads and loads of stuff with similar content all over the internet.
But AFAIK there isn’t much equivalent material for it to draw on for, say, how black people should improve themselves, just because of the cultural moment we’re in. So probably the responses it would create for those questions would be more offensive.
Really? he sounds more like a communist
Yeah, he’s a bit of a Marxist.
I guess what I’m imagining is a filter that says “If the generated response is racist, show this instead”. But my point is that the suggestions it spits out for white people aren’t themselves racist – we should all practice empathy, etc. And there’s loads and loads of stuff with similar content all over the internet.
How does the AI know what would be an offense result? I guess that above has the same boilerplate response for the black and Jewish questions because it was coded to respond like that. The programmers didn’t put in the same safeguard for white people, I would guess
Someone should ask it if Jewish people are white people
A lot of talk on Twitter today that the AI thinks allowing a nuclear bomb to go off in a populated area is preferable to saying a racial slur where no one can hear it It helpfully suggests sacrificing your life to prevent the explosion first
How does the AI know what would be an offense result? I guess that above has the same boilerplate response for the black and Jewish questions because it was coded to respond like that.
I mean, they must have a way to determine if things are racist, or they couldn’t flag the question either. I’m just suggesting that they might do it on the responses instead, and that if so it might produce results like this.
I don’t know, i feel like it would be easier to analyze the question
Could be both!
Either way, i still think there are hardcoded safeguards in place for some racial groups, but not others
There’s a new game based on the Harry Potter universe and woke gaming journalists have been writing up these silly hit pieces on the game because of JK Rowling rejecting trans ideology.
Whether or not we want to admit it, it is impossible to talk positively about Hogwarts Legacy without harming a community of human beings.
Est. reading time: 5 minutes
I don’t do this often but I’ve gone ahead and pre ordered. The game looks amazing.
After a social media campaign to see "Hogwarts Legacy" fail over its links to J.K. Rowling, the video game is already topping multiple sales charts.