Teach children to read with a legally-distinct cat in a legally-distinct hat! Leverage the power of machine learning / A.I. / other buzzwords!

Proposal:

Illustrated children’s books usually consist of two main things:

Some easy-to-read words
An illustration that is (hopefully) of interest to a child

Unfortunately, books written before the advent of computer-generated images are stuck with “stale” illustrations: each page contains a single fixed image that will never change.

But thanks to machine learning, we can fix this! Observe the new style of children’s book, where instead of an illustration, a QR code is present that links to an image generation prompt (Fig. 1a).

Fig. 1a: Instead of an illustration, this book contains an image-generation prompt (as a QR code). A child can just scan it to generate an image on the fly! (Note: this QR code just says “cat wearing a hat,” so don’t expect great things).

By changing the random number “seed,” a slightly different image can be created.

As an added bonus, it is very easy to reword a prompt if, for example, it turns out that the children’s book may have slightly promoted genocide / slavery / cannibalism / whatever (for a concrete example, see the revisions to the Oompa-Loompas from Charlie and the Chocolate Factory). In this fashion, a no-longer-socially-acceptable illustration can be redone in seconds by AI, rather than requiring complete re-illustration.

Example 1: “A Legally-Distinct Cat in a Hat”

Fig. 2: Here, we see an attempt at ripping off “The Cat and the Hat” (Dr. Seuss, 1957). The “style of Dr. Seuss” in the prompt has been blanked out (magenta rectangles, above) to craftily evade accusations of artistic theft.

PROMPT: a black cartoon cat wearing a red and white striped hat, a children’s book illustration, style of dr seuss, dr seuss, seuss, illustration, children’s book, crayon

NEGATIVE PROMPT: realistic

Example 2: “Goodnight To A Celestial Body Orbiting Earth”

Fig. 2: Even Goodnight Moon (Margaret Wise Brown, 1947) cannot escape the power of MACHINE LEARNING.

PROMPT: a cartoon green room with a red floor, fireplace, window with a blue sky and the moon behind, red window frame, a children’s book illustration, style of dr seuss*, dr seuss*, seuss*, illustration, children’s book, crayon

NEGATIVE PROMPT: realistic

[*] This book was not illustrated by Dr. Seuss, but the prompt includes that term anyway to get a flat cartoony style.

Example 3: “Two Fish, One Fish, Blue Fish, Red Fish”

∫∫∫

Fig. 3: Another example (arguably slightly less successful than the preceding two). This one is a ripoff of “One Fish, Two Fish, Red Fish, Blue Fish” (1960).

PROMPT: a white cartoon fish, two green cartoon fish, one cartoon red fish, one cartoon blue fish, yellow background, ocean waves at bottom, a children’s book illustration, style of dr seuss, dr seuss, seuss, illustration, children’s book, crayon

NEGATIVE PROMPT: realistic

Conclusion:

This is the future of art!

PROS: Children will love it! Adults may also appreciate the increased variety of illustrations.

CONS: Requires that the reader have a cell phone or VR headset. Presumably this will be the default in the future, so it’s not really a downside at all. Maybe VR headsets of the future will just seamlessly replace all QR codes with an image anyway, saving the user from having to manually input them!

Originally published 2024-11-18.