A blog related to original AI artworks. I make AI artworks using coined words and emoji as prompts.

“Ambiguous” Figures Created by AI Art

AI art software, especially Stable Diffusion 1.5 and OpenJourney, often depict “ambiguous figures.” In other words, they create images where certain parts or the entire form of a person are vague, yet it is still apparent that a person has been depicted. Faces are usually relatively clear, allowing us to recognize them as human. However, there are also cases where faces are blurry, yet still recognizable as human. The reason why such ambiguous paintings are created is likely because AI is not given a specific theme. However, one of the fascinating aspects of AI art is its ability to create images that painters rarely depict, like this one. In addition, Stable Diffusion 2.1 rarely depicts ambiguous figures like this.

First, let's look at examples from Stable Diffusion 1.5 where faces are relatively clear.

In the following images, although the faces are not clear, we can still perceive them as human.

When the entire body is depicted, even if it is vague, it seems easier to recognize them as human.

However, it is not clear how many people are present in the images.

The “ambiguous figures” produced by OpenJourney have a different atmosphere compared to Stable Diffusion.