Google's new "Storybook" AI tool creates illustrated stories, but with quirky results
The feature lets users select specific art styles—including claymation, anime, and comics—and even upload their own images for the AI to reference. Gemini can then create a story with short paragraphs of text and accompanying illustrations, which it can also read aloud.
ADVERTISEMENT
Google has launched a new "Storybook" feature within its Gemini AI chatbot, allowing users to create 10-page illustrated stories simply by describing them. While the tool offers a promising way to generate customized content for children, it's already showing some of the bizarre quirks often associated with AI-generated art.
The feature lets users select specific art styles—including claymation, anime, and comics—and even upload their own images for the AI to reference. Gemini can then create a story with short paragraphs of text and accompanying illustrations, which it can also read aloud.
However, a closer look at the stories reveals some strange inconsistencies. One story about a catfish in an aquarium generated an image of a fish with a human arm. Other examples include an AI-generated image of spaghetti sauce that resembled a "cartoon crime scene" and a TV screen shown on the wrong side in a picture of a mother and son.
Even in Google’s own promotional video for the feature, an AI image shows a woman "building a spaceship" while holding a wrench and making tapping noises, a scene described as having an odd, illogical quality.
While not all stories produced such blatant errors, the tool also struggled with a user's artistic vision, failing to accurately replicate a cartoon cat drawing uploaded for reference.
