ComradeSharkfucker, because I love you, I went and found these AI images from the brief period when Bing hardcoded “Ethnically Ambiguous” into its prompt field
I have genuinely checked that comment 2 times for an update since you posted it. I am so touched 🥺
Thank you comrade, I needed to see these lmao
Of course, I just got a little busy and then my partner was like “You’re working too hard, let’s get you something nice” and then we went and got strawberry lemonade sorbet from the nice ice cream shop
Then I realized how much time had passed and then I felt bad, so I found them and now i’m laying in bed curled up with my stuffed Trubbish and I may have also taken an edible or two
So in conclusion, I am glad you’re happy
Hell yeah, that sounds like a wonderful day. I’m laying on the couch with my partner watching avatar and I am about to go smoke lol. Enjoy cuddles with your trubbish ❤️
The lady growing out of the sidetable scares me
Oh yeah
I remember when nearly every genAI image was somewhat terrifying
Especially those ones where it would put eyeballs everywhere
The baby in the bottom right is horrifying
Actual photo of me
That’s right, I’m ethnically anbigrious
- I love our emojis
Etthnicicillly
AMBBIGOUSM
ambogus
If it’s hard coded in why does it do the text but not even spell it right but it “misspells” it in a consistent way
Generative AI stuff doesn’t have any concept of what a word is. It’s just trying to jam stuff together that looks vaguely like what its weird black box algorithm is telling it the result of the prompt should look like. And it’s not consistent at all lol, these pictures have ethnically anbigrious, ethinically ambigaus, ethnically ambigauus, etthnicicllly anbiguauu, and ethnniclly ambigaus
They trained the early models on whatever images they found on the internet. People noticed that “doctor” made a white dude and “basketball player” made a black dude, and made a stink about the racism in this*. Microsoft went into cover-my-ass mode and put in a simple AI model that checks for if the prompt sounds like it’s about people, and if it does, inserting “ethnically ambiguous” into the prompt before passing it to the image model**. The image models were shit, so they’d leak random parts of their prompt into any text on screen all the time.
* note that this wasn’t some intentional racist thing, it’s just a model thoughtlessly replicating the racism in its training data, after being made by thoughtless people who didn’t foresee this problem (ie: exactly how racism usually works)
** IIRC, google’s AI did the same thing, but instead of always adding “ethnically ambiguous”, it had some percent chance of adding a random racial descriptor. So you’d ask for George Washington and get a 10% chance it made him black
i’m not an expert with this stuff i’m afraid
Just telling you what I heard when it first happened
That Simpsons one is flat out racist.
Gen’AI’ used to be funny. Bad, but funny
Now it’s just women with creases and exploding dogs
ethical ambergris
Please add a content warning for racism to this post. Thanks.
Not going to remove it since it’s already falling off the front page and I know you’re asleep, but please be more mindful in the future. Even mockery of racism can unintentionally promote real racism if we aren’t vigilant.
Yup, sorry about that
Will be better about tagging things with CW’s
Ethnically babbroom
