normally, AI generates images by taking the prompt you give it, finding styles and similarities among previous-collected prompts and existing content, then combines various parts of content to produce a unified bit of artwork. Criticism of COMPAS highlighted that device learning versions are designed to generate "predictions" that happen to be