Design a Playground for One Specific Animal

A slide is useless to an octopus. Working out what would suit one instead means finding out how an octopus moves, and your child will do that research happily because it is in service of a playground.

How it goes

1. Pick one animal and stick with it. Odd choices work best: octopus, sloth, mole, flamingo.
2. Find out how it moves before designing anything. A book, a nature site, or a chatbot with the answers checked afterwards.
3. List five facts that change the design. Eight arms. Sleeps upside down. Cannot see well. Each fact rules something out.
4. Design the equipment, one piece per fact, drawn and labelled with the fact it serves.
5. Ask the AI to play the animal: You are a sloth. I will describe playground equipment and you tell me whether you could use it and why not. Being told “I move too slowly for that” by a sloth lands better than being told by a parent.
6. Redraw the pieces that failed.

Who does what

AI plays the animal, supplies facts to be checked, and asks what happens when the animal is tired or scared.

Your child researches, designs against constraints, defends each piece, and redraws.

You find the book so there is something to check the AI against.

Make it easier or harder

For a five-year-old: pick a familiar animal and two facts rather than five. A dog that likes digging and cannot climb gives you a sandpit and a ramp, which is a complete design at that age.

For an eleven-year-old: design for two animals that share the space and have incompatible needs, an owl and a rabbit. Working out what has to be separated, and what can be shared at different times of day, is a genuine constraint problem and it produces a much better drawing.

Take it off the screen

Build one piece of equipment at model scale from straws, card, and string, with a plasticine animal to test it.

Then visit a real playground and ask which pieces would work for the sloth. Children who have done this look at climbing frames differently for months.

When it goes sideways

If the AI plays the animal too obligingly and approves everything, sharpen the instruction: Refuse anything that does not match how a sloth actually moves, and say which fact it breaks. Agreeable roleplay is useless here; the value is in being told no.

If your child designs a human playground and paints the animal on it, go back to the five facts and ask which piece serves which one. A piece that serves no fact comes out. That single question fixes the project almost every time.

One thing to watch

Animal facts are exactly where a model produces confident errors, and your child is designing around those facts. Check the five that matter in a book or on a site that names its sources. Seek is good for identification if you want to start with an animal in your own garden.

Related: invent a mythical animal is the fictional version, and design a future city scales the same design-against-constraints idea up.