DALL·E 3 is an advanced text-to-image system that generates highly accurate images based on the text provided. It is an improvement over previous versions, allowing users to easily translate their ideas into detailed and nuanced images. It is built natively on ChatGPT, which enables users to interact with the system and refine their image prompts. DALL·E 3 focuses on safety by limiting the generation of violent, adult, or hateful content. It also includes mitigations to decline requests for public figures and addresses biases related to visual representation. The tool provides creative control to users and allows them to opt their images out from future training.