Here’s exactly how to do it
Imagine if you could take your child’s scribbles and transform their ideas into artist-quality illustrations. With generative AI, that’s now not only possible — it’s easy and cheap.
Earlier this month, OpenAI released two new features in ChatGPT. ChatGPT Vision allows the popular AI system to analyze the content of an image, and DALL-E3 allows is to create wholly new images all by itself.
I’ve been using these systems for lots of complex work stuff, like analyzing charts in government reports and extracting prices from restaurant menus.
But I’ve also been using it with my kids. And I recently found an incredibly cool way to combine these two new tools together.
My three-year-old son recently took a marker and scribbled out a little drawing, as three-year-olds do.
Like many preschoolers, his imagination went far beyond what his fine motor skills can manage. When I asked about his drawing, he explained that it was “a little triangle car with a smiley face and a rainbow.”
I decided to see if ChatGPT could bring his ideas to reality. First, I uploaded a cellphone photo of his drawing, and gave ChatGPT Vision the following prompt:
Here is a child’s drawing of a car. Describe it in enough detail that an illustrator could transform this idea into a detailed drawing
Here’s the response:
The drawing depicts a somewhat abstract representation of a car.
The car’s main body appears triangular or cone-shaped, with the wider base at the bottom. The body has a smiley face drawn on it, which serves as the front of the car.
Extending from the top of the triangular body is a single, wavy line that could be interpreted as an antenna or a quirky design element.
On either side of the triangular body are what appear to be wheels. The wheels are not perfect circles but rather are squiggly, almost…
