Using a dataset of text-image pairs, the 12-billion parameter DALLE variant of GPT-3 was trained to produce images from text descriptions. It is capable of a wide range of tasks, such as anthropomorphizing objects and animals, convincingly merging seemingly unrelated ideas, producing text, and altering already-existing visuals.
https://openai.com/blog/dall-e/
DALL-E Image generator webpart by Luis Manez
https://github.com/pnp/sp-dev-fx-webparts/tree/main/samples/react-dall-e-image-generation