What's your personal AI code of conduct? Every kid already has one. Most never wrote it down.

Ask a teenager what their AI code of conduct is and you will get a blank stare, followed by something diplomatic about not cheating. But watch what they actually do at eleven p.m. with an essay due at midnight, and you will discover they absolutely have a code. It was just never written down, never chosen, and never read aloud in front of anyone they respect.

The unwritten code usually has one rule: whatever gets this off my plate. It is a perfectly rational code for a kid who has been told AI is either forbidden magic or free homework. Nobody handed them a third option.

Here is our claim, and it is the hill FutureMasters is built on: the single highest-leverage thing a young person can do with AI right now is not learn a tool, a trick, or a prompt. It is to sit down and write their own rules for how they will and will not use the most powerful technology of their lifetime. On purpose. In their own words.

Put two kids in front of the same model and watch the codes collide. The copy-paster opens a chat, pastes the assignment, copies the answer, and closes the tab. Total thinking time: the walk to the fridge while it generates. Ask them to defend a sentence of it tomorrow and they will look at you like you asked them to defend a stranger's diary.

The high-agency kid opens the same chat and does something that looks almost rude to the machine: they argue with it. They bring their own draft, their own half-formed take, and make the model attack it. They ask for three angles they had not considered, take one, and throw the other two away with prejudice. They use it to check their logic, not to replace the part of them that has logic.

The cruel arithmetic is that both kids spent the same hour. One of them compounded. The copy-paster's skills are exactly where they were in September, except now with a dependency. The high-agency kid got faster, sharper, and harder to fool, because their code routed the machine's power through their own judgment instead of around it.

So what goes in a personal code? Not commandments from adults. Kids can smell imported morality from across the house, and a rule they did not author dissolves at the first deadline. The code has to be theirs, which is exactly why it works. Nobody defends a fence they did not build.

The strong codes we see kids write share a spine, four clauses long:

Notice that none of these rules are about the technology. They are about sovereignty. Which is why a kid's AI code of conduct is really a first draft of their character, written in the one arena where the temptation to outsource themselves is strongest.

Our favorite clause deserves its own act: the sparring partner rule. The weakest possible use of an intelligence greater than yours is asking it to do your reps. The strongest is making it your opponent. Have it grade your argument like a hostile examiner. Have it find the flaw in your plan before your audience does. Have it defend the opposite position at full strength and see if you survive.

A kid running this clause gets something school rarely provides: unlimited, patient, ego-free opposition. They can be wrong in private a hundred times an hour, which is the fastest known way to end up right in public. The machine takes no offense and gives no mercy, and the kid's ideas come out of that room with muscle on them.

This is also the cleanest answer to the fear that AI will make kids soft. A punching bag never made anyone weaker. It depends entirely on whether you are throwing the punches or asking the bag to throw them for you.

Parents, here is your field guide, because from the hallway the two codes can look identical: a kid, a screen, a chat window. Support loudly when you see these: your kid explaining what the AI got wrong; drafts that existed before the machine was opened; your kid disclosing AI help without being cornered; questions getting stranger and more ambitious over time; projects that exist because the machine made them possible, not because it made them free.

Discourage without apology when you see these: answers your kid cannot re-derive out loud; the blank-page reflex, where the chat opens before the thinking starts; secrecy about what was generated; and the saddest tell of all, output getting more polished while your kid gets less interesting.

But the most powerful move a parent can make is not surveillance. It is one dinner-table question: what are your rules for this thing? A kid who has an answer is already ahead of most adults. A kid who does not will happily build one if someone they respect acts like it matters. That conversation costs nothing and quietly settles the only question that counts: who is operating whom.

At FutureMasters, writing a personal AI code of conduct is week-one work, before any tool touches a project. Every kid drafts their rules, defends them to a room of peers, and then spends twelve weeks getting caught between their code and their deadlines, which is where character actually gets built. The code gets revised. It does not get abandoned, because it is signed with the only signature that binds a teenager: their own.

The machines are not going to get less capable, and midnight is not going to get less tempting. The kids who thrive will not be the ones with the cleverest prompts. They will be the ones who decided, early and in writing, what they refuse to outsource.

So ask your kid the question in this page's title. Then, and this is the hard part, ask yourself. The best code of conduct in the house is still the one running in front of the kids.