A parent said something to us recently that we have heard in one form or another almost every week since. “I was going to sign him up for coding, but honestly, what is the point now?” It is a fair question, asked in good faith, and it deserves a straight answer rather than reassurance.

The short version is this. AI has made the production of answers cheap. It has not made the judging of answers cheap, and it never will, because judging requires knowing what good looks like. A child who has never had to think their way through a hard problem has no idea what good looks like. That is the gap, and chess and coding happen to be two of the best ways to close it.

1The question parents are quietly asking

The real worry underneath “is coding still worth it” is usually broader: am I preparing my child for a world I cannot picture? That anxiety is reasonable and it is not new, but it has sharpened.

It helps to separate two different things. One is teaching a child a tool that may be obsolete in a decade. The other is teaching a child a way of thinking that outlives whatever tool it was learned through. Handwriting did not stop being worth teaching when word processors arrived, because the point of handwriting was never the letters. Similarly, the point of a coding class for a ten year old was never that they would be employed writing that language.

2What AI is good at, and what it is not

AI is extremely good at producing something plausible very quickly, and noticeably weaker at knowing when it is wrong. That single asymmetry explains most of what parents need to understand.

Think about what actually happens when a child asks an AI to solve a maths problem. They get a clean, confident, well-formatted answer in two seconds. If the answer is right, they learned nothing about how to get it. If the answer is wrong, they have no way to tell. Both outcomes leave the child in the same place: dependent on a tool they cannot check. The child who can check is in a completely different position, and checking is a skill built by struggling with problems that had no shortcut available.

This is not an argument for keeping children away from AI. They will use it, and they should. It is an argument for making sure something in their week requires them to think without it.

3What employers actually say they need

Employers are not asking for fewer thinkers. In the World Economic Forum’s Future of Jobs Report 2025, analytical thinking remains the top core skill, named as essential by around seven in ten companies surveyed, with resilience, flexibility and agility close behind. AI and big data are indeed the fastest growing skills, but they sit alongside creative thinking, curiosity and lifelong learning rather than replacing them.

The number parents should sit with, though, is a different one. The same research finds that employers expect roughly 39% of required skills to change by 2030. If two fifths of the specific skills turn over in five years, betting your child’s education on any particular tool is a poor bet. Betting it on the capacity to pick up new tools quickly is a much better one.

4Chess as a thinking gym

Chess earns its place because it forces a child to make a decision under uncertainty and then face the consequence immediately, with nobody to blame. There is no group to hide in and no teacher marking it later. The board tells the truth within about four moves.

That feedback loop is the whole value. A child considers options, picks one, discovers it was wrong, and has to keep playing anyway. Over a few hundred games this quietly teaches something no worksheet does: that being wrong is information, not identity. It is also the reason chess suits anxious and perfectionist children so well, provided the adults around them handle losses calmly.

Be honest about the limits, though. The research is encouraging rather than spectacular. A review of chess instruction studies by Sala, Foley and Gobet found small to moderate improvements in mathematics and general cognitive measures, with the effect growing as training hours increased, and the authors themselves flag that many studies lacked strong controls. Chess is a good use of a child’s hours. It is not a magic pill, and any tutor who promises you one is overselling.

5Why coding still matters when machines write code

Coding is worth learning because the skill was never typing. The skill is breaking a vague goal into exact, ordered, unambiguous steps, and that is precisely what a person needs in order to direct an AI usefully rather than just accept whatever it produces.

There is a practical version of this that every parent will recognise. Ask an adult to instruct an AI and the ones who get good results are the ones who can specify clearly what they want, notice when the output is subtly wrong, and describe the correction. Those three abilities are exactly what a child practises when their program runs but produces the wrong shape on screen and they have to work out why. A meta-analysis of coding interventions with children found the largest effects were on problem solving specifically, which is consistent with what coding teachers see: the debugging is the lesson.

For a young child, the language barely matters. Block based tools such as Scratch teach sequence, loops and conditions without typing. Text based languages come later. What carries forward from either is the habit of asking “what exactly did I tell it to do, and how is that different from what I meant?”

6The skill underneath both: judgement

Strip chess and coding back and the same thing sits underneath both: judgement, which is the ability to evaluate options that all look reasonable and commit to one. That is the capacity AI cannot hand to a child, because judgement is built by making calls and being wrong, not by reading correct answers.

Notice how similar the two activities are once you look past the surface. In chess, several moves look fine and the child must choose. In coding, several approaches would work and the child must pick one, then find out where it breaks. In both cases the child is the one who decides, the feedback is immediate and impersonal, and the cost of being wrong is small enough to be survivable. That combination is surprisingly rare in a child’s week. School mostly asks for the right answer. Sport asks for effort. Very few activities ask a child to weigh, choose and own it.

7How to tell whether an activity builds thinking or replaces it

There is one reliable test: does the activity make your child produce something, or consume something? Producing builds judgement. Consuming, however educational the label, generally does not.

Some practical markers. Does the child have to make choices, or follow steps? A class where every child ends up with the identical output is a tutorial, not a thinking exercise. Does the child get told they are wrong quickly? Fast, low-stakes failure is the engine of learning, and an activity that never lets a child fail cannot teach them to recover. Can the child explain what they did afterwards? If they cannot narrate their own reasoning, the reasoning probably was not theirs.

Apply that test to AI use as well. A child using AI to explain a concept they then apply themselves is producing. A child using AI to hand in the finished thing is consuming, and no amount of technology literacy makes that a good trade.

8What this means for choosing your child’s classes

Choose the class where your child has to think out loud in front of someone who is paying attention. In practice that usually means a small setting or a personal one, because thinking out loud is exactly what gets skipped when a teacher has fifteen children and a syllabus.

The questions worth asking a prospective tutor are simple. What happens when my child gives a wrong answer? If the answer is “we correct them and move on,” look elsewhere. The valuable response is a question back: why did you choose that? What else did you consider? That single habit, repeated weekly, is the whole of what we are talking about in this article. It is also, frankly, the main reason 1 on 1 works better than a group for this kind of learning. There is nowhere for a child to be quiet and get away with it.

And do not overthink the subject choice. Chess, coding, maths, even a musical instrument taught well: what matters is whether the child is deciding things. Pick the one your child will still be interested in at week twelve, because consistency beats subject every time.

Where to start

Start with one activity, once a week, where your child has to decide something and find out whether they were right. That is the whole prescription. It does not have to be chess and it does not have to be coding, though both happen to be unusually good at it.

If you would like to see how your child handles that kind of thinking before committing to anything, a free trial class is the least risky way to look. You will sit in, watch your child reason out loud for fifty minutes, and know a great deal more than you did an hour earlier. There is no cost and no obligation to continue.