Why is it so flawed?

"AI lacks emotions that creates art – AI is motivated by commands, not a desire to express itself. Works are created with no intent and no sense of what’s relevant. Humans are required to interpret AI's outputs."

"AI art lacks a dialogue with other values that inspire artists – AI generates imagery based on a dataset that does not include how memories, symbols, language and cultures influence our own organic neural network – the brain."

"It can't come up with new ideas – content created by AI is based on existing data it has been fed and cannot think up new ideas without new inputs."

(University of Plymouth, 2023)

What is wrong with the hair? Is that supposed to be a mermaid tail?

In short, AI doesn't know how art nor the world works. How can that be? AI knows as much as the creator feeds it. When a programmer trains AI on what they find on the internet, there is bound to be inconsistency. Just as no two people are exactly the same, no two art pieces are the same. With each artist having their own art-style portrayed in their work, AI generates its understanding of patterns, patterns which appear inconsistent when recycled in an image.

The image above is an example of this confusion in pattern recognition. One of the first features that you notice in the bold orange hair, The hair has no set direction it is going and it is visible where the fed data patterns contradicted itself. One dataset tells the AI that the hair is more curly and ends at the back, while the other tells the AI that the hair must flow long past the body, flowing but straight. There is also the top of the head that appears to be a different material than hair completely.

Another strong example is what appears to be the tail of the mermaid. The AI used to generate this image appears to not have an understanding on what a mermaid is. It is understood that "a mermaid" is typically depicted as female, in water, minimal upper-body clothing and that it does not have human legs. However, that appears to be the extent of the AI's training with the tail being part scaly with the rest being a smooth multi-colored geometric tube that ends with a rectangular fin-like shape. There are also unidentifies growths pertuding from the tail, adding onto the difficult to understand shape.

There are many more visual distortions in the presented image, such as the moon floating in the water and the arm. However, smaller flaws such as these can often be overlooked by those who are not looking. When AI art is generated, there tends to be both major and minor inconsistencies as a result of the AI not understanding the world and attempting to learn by categorizing art made by a large variety of non-consenting people to collage and distort to fulfill its command.