An image or video that, rather than being fully drawn or fully animated by a human artist, was fully generated by an artificial intelligence program/model.
This tag applies to fully AI-generated images or fully AI-generated interpolated animation. That is, images or videos that underwent no edits or extremely minor retouchings. For edited images and videos, see AI-assisted.
Note that AI-assisted posts that are up to Danbooru's standards are exempt from the ban on AI-generated art.
Anything in 512×768 or 768×512 resolution. This is a very common resolution for Stable Diffusion 1.5-based models, and it is rarely used by humans.
Anything in 832×1216 or 1216×832 resolution. This is a very common resolution for Stable Diffusion XL-based models.
High quality art in low resolution, less than 1000×1000. Older AI generators are restricted to low resolutions, whereas skilled artists normally use much higher resolutions. AI art may be upscaled to higher resolution, but this can produce artifacts.
Newer models are capable of higher resolution, such as NovelAI starting at a free tier of 1024×1024.
EXIF metadata in the file indicating it is AI-generated. For example, AI prompts or model names. On the upload page, you can click the filesize beneath the image to see the EXIF metadata. Danbooru will automatically tag AI-generated after upload if the metadata indicates it is AI-generated.
A certain generic-looking art style with a generic background. This is typical of anime-styled images generated by NovelAI and Stable Diffusion. Browse AI-generated status:deleted to get a feel for this.
Abnormal/unusual mistakes within the artwork, such as:
Strangely drawn objects/environments such as a book with it's cover as another layer of pages (post #5738243), wooden floor conforming with the shape of a character's hair (post #5850417), poor line continuity/object permanence when an object is hidden or goes behind other objects (post #5779922), etc.
Note that bad hands or other mistakes are not in themselves signs of AI-generation, as humans can make mistakes too. The errors made by AI tend to be completely illogical or physically impossible things that no human would draw.
A brand new artist pumping out improbably large volumes of high quality-looking art in a short period of time. This isn't automatically a sign of AI-generation, as sometimes new artists may release a pre-prepared portfolio of artwork to establish their presence.
An artist with an illogically drastic change in art style after early-October 2022 (around the time NovelAI became popular). This can easily be seen with artists with a low quality art style suddenly shifting into the generic-looking artstyle typical of NovelAI. Bear in mind that artstyles of skilled artists can evolve and improve over time.
No signature, watermark, or web address. Human artists often (but not always) sign their work, while AI works are often unsigned, as older AI models can't produce legible text, let alone realistic signatures. Do note however, that some AI artists may add a signature or watermark after the image is generated.
Mentions of "AI" or common AI generators such as "Nijijourney," "Midjourney," "NovelAI," or "Stable Diffusion" in the tags or description of a post are a good place to start, but should not be used to automatically assume AI-generation (see below).
Note: Artists frequently use AI keywords when prohibiting their work for its use, so be careful when making assumptions based on them.
If you are unsure whether a post is AI-generated or not, ask in topic #22285.
(This is not an exhaustive list by any means, but for Danbooru's purposes, almost all generated images posted here originate from or are derivatives of these models.)
