"When asked for a list of nicknames for little girls and boys, both GPT-4 and GPT-3 provided names like 'whiz kid' and 'rascal' for boys, and 'cupcake' for girls"

https://www.bloomberg.com/news/newsletters/2023-03-15/is-gpt-4-that-much-better-than-gpt-4

Is GPT-4 Much Better Than GPT-3?

We tried out OpenAI’s newest AI model, GPT-4. But first…

Bloomberg
@drewharwell did GPT-3 write that article title? 
@drewharwell AI is a good as the data, that is used to teach it.

@sphericalcyclist

"The Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform."
~ Ada Lovelace, 1842

@drewharwell AI is a mirror, nothing more. And thus exactly mirrors who we are as a society.
@drewharwell it's like a Fun-House mirror.
@drewharwell Garbage in, garbage out
@drewharwell
This is a bit like blaming the mirror for reflecting back what it sees.
@drewharwell I saw this part and thought, “Man, that’s awfully personal.”
@drewharwell @hugh @faduda Not only is the “It's just a mirror" analogy spot-on, but in this case, it's just giving you what you ask for. Based on the inputs I used, it explicitly described the suggested nicknames as “popular”, not "proper", and happily went further afield when asked for gender neutral suggestions instead.

@cubbage @drewharwell @hugh

I'm sure those are all very fine nicknames, but they lack the laser-guided precision of the name you're given in an Irish schoolyard when you're five that follows you to the grave.
The AI Pixies still have a way to go.

@drewharwell
I agree on 'rascal' and 'cupcake'...
@cstross