The Drawbacks with AI Illustrations

One commitment we can make to our readers and supporters is that we have not and will not use artificial intelligence (AI) in our books. Our mission is to provide diverse books for all readers written and illustrated by diverse people. People, not AI.

However, we know that many authors who are desperate to get their manuscripts into book format and into the market see AI as a shortcut because it is fast and cheap. There are several reasons we believe people should reconsider.

AI beauty standards are based on warped human beauty standards.

Dove, a company that makes products for skin and hair, recently released a short video addressing the beauty images that one might see when using AI. Those images were of white females with blond hair and blue eyes. This line appears at the end of the video: “Dove will never use AI to create or distort women’s images.”

In its commercials, Dove has been intentional about using people of all shades of skin tone, hair texture, age, and body size. When people choose to use AI for books, they may end up with images that are distorted such as an ethnic person with features that seem more European.

AI pulls from existing beauty norms. When working with human illustrators, those standards can be obliterated, and people who don’t tend to see themselves in book can now see themselves.

AI is Inconsistent

One important aspect of illustrations is consistency. Characters should look the same on every page. When AI generates a book, that consistency is not there and it is not able to pinpoint issues and only target the issue area. In a Time Magazine article about an AI created book, several issues were noted about the illustrations.

The AI-generated illustrations had a number of issues: some fingers looked like claws, objects were floating, and the shadowing was off in some areas. Normally, illustrations in children’s books go through several rounds of revisions—but that’s not always possible with AI-generated artwork on Midjourney, where users type a series of words and the bot spits back an image seconds later.

Some people try to get around this by making books where each page is a biography about a different person. That gets the person around the issue of the character looking the same throughout the book since the character is only depicted once. However, that doesn’t get the person around the inconsistency between characters. There are some AI illustrated books where some characters have cartoon illustrated noses and others have more human-like illustrated noses within the same book. Think of a book like a universe. In the universe the style of illustration should be the same. Some characters should not look nearly life-like while others look like Saturday morning cartoons.

AI generates images based on real art

Adobe Firefly defines AI art as  “artwork made with the assistance of generative AI — a technology that finds patterns in big datasets and uses that information to create new content.” In this case a picture or piece of existing art may be submitted into an AI platform and enhanced. Additionally, people can generate “new” images from adding a few key words. However, those images are based on images already in existence. That includes work from people who make a living as illustrators.

Many artists are part of the #NoToAIArt campaign. In a Guardian article, illustrator Anoosha Syed said, “AI doesn’t look at art and create its own. It samples everyone’s then mashes it into something else.”

In short, support illustrators. Don’t take short cuts. The full legal ramifications of AI art are still unknown. Good books take time. Trust the process. Trust your illustrators. Support human artists who are doing the work that AI steals to create “art.”