Will starting with AI make my kid skip "real" coding?

No — directing an AI exposes the code and the logic. Kids who want to go deeper have a reason to, which is the opposite of skipping it.

My child is 7 — is that too young for either?

Not for guided, plain-language building. Open-ended chatbots aren't built for that age, but a project-based tool with guardrails is a good on-ramp.

Should kids learn coding before AI, or AI before coding?

For most school-age kids, start with a project built by directing an AI, then move into the underlying code when the child wants control the plain-language route will not give them. The two are one skill, so the practical question is only which entry point gets a real project finished fastest — and for this age that is almost always the AI-directed one.

How long should the first stage last before we look at real code?

Usually three or four projects, not a fixed number of weeks. The signal is the child expressing a specific frustration — "I want it to do X and it keeps doing Y." That frustration is the readiness cue, and waiting for it means the code arrives as an answer rather than as an assignment.

What if my child's school teaches Scratch already?

Then the fundamentals are being covered and you can use home time for the part school rarely has room for: a self-chosen project, and the habit of checking and explaining. Duplicating the school's curriculum at home is the least useful thing you can do with those hours.

Does the order matter if my child is 12?

Less. At twelve the syntax tax is smaller and abstraction is easier, so a motivated child can start from either end successfully. The order matters most at seven to nine, where the gap between "can hold an idea" and "can type it correctly" is widest.

Parent decision guide

Coding vs AI for kids: which should they learn first?

Short answer: they're one skill, so start with whichever gives your child a real project fastest — and for most kids that's building something by directing an AI. You get coding thinking and AI literacy in the same activity, and the "which first" worry disappears.

If you want a concrete order

Begin with a small, motivating build — a simple game or a story tool — where the child directs an AI in plain language. That teaches precise instruction, testing, and cause-and-effect immediately. When your child hits something they want more control over, that's the natural moment to go deeper into the underlying code (Scratch or Python). The project pulls the fundamentals in, instead of the fundamentals gatekeeping the project.

The reason order matters at all is attrition. Most children who quit coding quit in the first six weeks, and they quit during the part where nothing they care about has happened yet. Any sequence that front-loads a working project and back-loads the fundamentals will keep more kids in the subject than the reverse, even if the reverse looks more rigorous on paper.

The order, concretely

1. 1. One project, chosen by the child Not assigned, not from a list of three. The child names a thing they want to exist. It will be too ambitious; that is fine and even useful, because scoping it down together is itself the first real engineering conversation.
2. 2. Direct, run, and see it fail The first version will be wrong in some visible way. Do not skip past this. The gap between "what I asked for" and "what I got" is the entire lesson in precise instruction, and it lands far harder when the child discovers it than when an adult explains it.
3. 3. Refine the instruction, not the code Before touching anything technical, have them re-say what they wanted more exactly. Most first-round failures are specification failures. Kids who learn to notice this become dramatically better at directing any tool, and better at asking teachers and adults for what they need.
4. 4. Open the code when curiosity arrives Somewhere around the third or fourth project, a child will want a change the plain-language route handles clumsily. That is the moment to look at what is actually running. Introduced then, code reads as power. Introduced in week one, it reads as an obstacle.
5. 5. Explain it to someone The project is not finished when it works; it is finished when the child can say how it works to a person who was not there. This is also the only reliable way for a parent to tell whether the learning happened.

The case for the opposite order — and where it holds

It is worth being fair to the traditional sequence, because it is not wrong for everyone. Starting with fundamentals — variables, loops, conditionals — before any AI involvement genuinely builds a sturdier mental model of what the machine is doing. A child who has written a loop by hand understands what the AI produced in a way a child who has not simply does not.

That order works well for a specific kind of child: one who is already motivated, patient with abstraction, and old enough that the syntax tax is small. If you have a persistent eleven-year-old who has asked to learn to code, starting in Python is a fine and defensible plan.

It works badly for everyone else, and "everyone else" is most 7–10 year-olds. For them the abstraction has no anchor yet, and six weeks of exercises before the first thing they care about is a reliable way to produce a child who concludes they are bad at this. The AI-first order is not better in principle — it is better at keeping a child in the room long enough for the principles to matter.

Signals you picked the wrong order

They can produce but not explain — The child has finished projects and cannot say how any of them work. The tool is doing the thinking. Slow down, shrink the projects, and require an explanation before anything counts as done.
They never get stuck — Frictionless is a warning sign, not a success metric. If nothing has broken in a month, the projects are below the child's level and no debugging — the most transferable skill here — is being practised.
They only work when told to — Unprompted tinkering is the single best indicator that a format has taken. Six weeks of compliance with zero self-started projects means the activity has not connected to anything the child wants.
They have stopped asking why — Early on, kids ask how the machine knew what to do. If those questions dry up, the child has settled into treating it as magic — the exact habit this whole sequence exists to prevent.

Quick answers

Will starting with AI make my kid skip "real" coding?: No — directing an AI exposes the code and the logic. Kids who want to go deeper have a reason to, which is the opposite of skipping it.
My child is 7 — is that too young for either?: Not for guided, plain-language building. Open-ended chatbots aren't built for that age, but a project-based tool with guardrails is a good on-ramp.
Should kids learn coding before AI, or AI before coding?: For most school-age kids, start with a project built by directing an AI, then move into the underlying code when the child wants control the plain-language route will not give them. The two are one skill, so the practical question is only which entry point gets a real project finished fastest — and for this age that is almost always the AI-directed one.
How long should the first stage last before we look at real code?: Usually three or four projects, not a fixed number of weeks. The signal is the child expressing a specific frustration — "I want it to do X and it keeps doing Y." That frustration is the readiness cue, and waiting for it means the code arrives as an answer rather than as an assignment.
What if my child's school teaches Scratch already?: Then the fundamentals are being covered and you can use home time for the part school rarely has room for: a self-chosen project, and the habit of checking and explaining. Duplicating the school's curriculum at home is the least useful thing you can do with those hours.
Does the order matter if my child is 12?: Less. At twelve the syntax tax is smaller and abstraction is easier, so a motivated child can start from either end successfully. The order matters most at seven to nine, where the gap between "can hold an idea" and "can type it correctly" is widest.

Give them one path that covers both

Luchi missions start with a project and teach the thinking behind it — no wrong first choice.

Get on the list

Back to the decision guide

Part of the series: Should my kid learn coding or AI?

Related guides in this series

Will AI replace coding? — AI changes what coding is, not whether your kid should learn it. The durable skill is telling a computer what you want — and how to check it.
Is coding still worth learning for kids? — Yes — but the skill shifted from syntax to thinking and directing. Exactly what a kid practices when they build with AI.
What age should kids start with AI? — A practical, age-by-age answer for 5–12s — when AI helps, when to wait, and what “starting” should actually look like.
Best AI apps for kids, compared — Eight products side by side, every claim quoted from the product’s own site with the capture date — including ours, and including where a cheaper option is the better buy.