“Dad, can you make me a dinosaur wearing a hat?”
This is how my 7-year-old asked me to use an AI image generator for the first time. He didn’t know it was AI. He didn’t care. He just wanted the dinosaur in the hat.
I made it for him. He laughed. Then — and this is the part that mattered — he said: “Can you make a purple dinosaur with two hats?” Then: “Can you make him eating ice cream?” Then: “Can the ice cream be on fire?”
In about 6 minutes my son went from passive viewer to active director. He wasn’t using AI. He was commanding it. The same kid who would otherwise have spent that 6 minutes scrolling YouTube was now art-directing a comic book in his head and using me as the rendering engine.
That’s the gap this article is about. Most kids using AI today are using it to get something — answers, summaries, jokes, distractions. The skill worth teaching is using AI to make something — an image, a story, a song, a project. Same tools. Completely different relationship.
Why “maker, not consumer” is the right frame
When a kid uses ChatGPT to write their book report, they’re a consumer. The work happens to them. They come away with a thing they didn’t make and a feeling that AI is magic they can’t really touch.
When a kid uses an AI image tool to design the cover for a story they wrote themselves, they’re a maker. The work happens through them. They come away with a thing they couldn’t have made alone, a clear understanding that AI does what they tell it, and the slightly addictive feeling of “what else could I do?”
Same kid. Same tool. Same hour. Two completely different developmental outcomes.
The maker frame builds two of the AI Readiness Framework dimensions at once. It builds Skills because the kid is learning to direct AI rather than be directed by it. And it builds Communication because every time they refine a prompt — “no, with the hat tilted, and make the dinosaur look surprised” — they’re practicing the rare skill of describing what they actually want to another mind.
The output (a dinosaur in a hat) is irrelevant. The reps are the point.
Four free tools to start with this weekend
I’ve tested all of these with kids ages 6-12. None require a credit card. All work in a browser.
1. ChatGPT (free tier) for stories. Have your kid invent a character — name, what they look like, one weird thing about them. Then together, ask ChatGPT to write a 3-paragraph adventure starring that character. The first version will be generic. The point is the second version: ask your kid “what should we change?” Get more specific. Iterate 3-5 times. By round 5 your kid will be giving direction like a producer (“make the dragon speak in rhymes, but only when he’s scared”). That’s the skill.
2. Microsoft Designer / Canva AI (free tier) for posters. Ask your kid to design a poster for an imaginary event — “Birthday party for a cat that loves space.” Give them the prompt box. Watch them try to describe what they want. Watch them be frustrated when the first try doesn’t match what they imagined. Don’t fix it for them. Ask: “what word should we change?” The frustration is the skill being built. After 4-5 iterations, they’ll have a poster they love and a real understanding that AI does what you say, not what you mean — which is the most important AI literacy skill there is.
3. Suno (free tier) for songs. Suno turns text into actual music with vocals. Have your kid write 4 lines about something they care about (a pet, a video game, a sibling). Pick a music style together — country? hip-hop? sea shanty? Generate. Then the next round: “what would make it better?” Faster? Sadder? Funnier? Within 10 minutes you’ll have a song your kid actually wrote — and your kid will have done more songwriting than 99% of children their age have ever done.
4. Scratch + AI extensions for projects. This one is for kids 9+. Scratch is the free MIT-built coding playground millions of kids use already. There are now AI extensions that let a Scratch project use AI to recognize voices, identify objects on a webcam, or generate text. The leap from “I made a cartoon character move” to “I made a cartoon character that listens to me” is magical and lasting. (If your kid hasn’t tried Scratch yet, this is the moment.)
How to actually do the session (the part nobody writes)
Here’s the thing nobody tells you about creative AI sessions with kids: the parent’s job is not to type.
The parent’s job is to ask three questions, on a loop:
- “What do you want?” — at the start
- “What should we change?” — after every output
- “Is it what you imagined?” — at the end
The kid does the typing. The kid makes the decisions. You’re a coach, not a co-author. If the kid says “make the dragon green,” you don’t override and say “no, blue is better.” You type “green dragon.” The dragon is bad? You ask “what should we change?” not “let me fix it.”
This is harder than it sounds. As a parent — and especially as a dad with engineering instincts — the urge to optimize the output is enormous. Resist it. The output is a byproduct. The skill being built is the kid learning to think out loud about what they want, which is the foundational skill of every form of creative work, with or without AI.
10 minutes a week of this beats 10 hours a month of you “showing them how to use it.”
Where to draw the lines (because there are lines)
Three things to be careful about:
Age limits. ChatGPT’s terms say 13+. Suno says 13+. Most AI tools have an age floor. For younger kids, you operate the account and your kid sits next to you and directs — same as you’d help a 7-year-old order at a restaurant in another language. This is also a great way to make sure you see what’s going in and what’s coming out.
The “is it real?” question. A creative AI session is also a great opportunity to ask: “Is the dinosaur in this picture a real dinosaur, or a made-up one?” This is where the Skills dimension of digital self-defense plugs in. AI tools will happily generate things that look like photos but aren’t. Your kid should know that distinction by age 8.
The “did you make it?” question. When your kid’s poster comes out perfect, the temptation to post it on a fridge or send it to grandma without context is real. Don’t. Always say “made by [kid’s name] with AI help.” The kid hears the credit. The kid also hears the disclosure. Both matter — and it sets the model for what intellectual honesty around AI looks like for the next 15 years.
Where this fits in the framework
The AI Readiness Framework has a Skills dimension that most parents read as “can my kid use AI?” The actual skill is direction — can your kid make AI do what they want, instead of accepting whatever AI gives them by default? Creative tools are the fastest way to build this dimension because they have an immediate visible payoff (the picture, the song, the poster) that motivates iteration in a way that homework never does.
It also builds Communication, because every prompt refinement is practice in describing what you want clearly enough that another mind — silicon or human — can act on it.
Take the AI Readiness Quiz → to see where your family stands across all 5 dimensions. If “Skills” is your low score, this article is your starting point.
Start this weekend
Pick one tool. Pick one project. Pick one kid. Sit down for 10 minutes. Ask “what do you want?” Type what they say. Don’t fix anything they don’t ask you to fix.
If they want to keep going past 10 minutes, you’ve succeeded — that’s the difference between a kid who consumes AI and a kid who commands it. Let them keep going. The dinosaur in a hat is a side effect. The maker is the point.
Want a guided 10-minute version? The First AI graphic challenge walks through one image-generation session step by step.