“DALL·E is a 12-billion parameter version of GPT-3 trained to generate images from text descriptions, using a dataset of text–image pairs.”
Fiddling around with DALL-E, I came up with these images. I haven’t learned how to prompt correctly, so none of them really work for my use case. But they are kinda pretty. My prompts were variations on “red-skinned female hellbred cleric wearing a black nurse’s cap comic book fantasy art.”
I noticed that DALL-E gave me white people until I asked for red skin, and then I got mostly black-skinned people. Apparently, the default is white, even in DALL-E.
