Where the Laws Stop Short

In the last few weeks, the adults moved on kids and AI from every direction at once.

New York's teachers union called for hard limits on student-facing AI, especially for the youngest grades. The state legislature passed a kids' chatbot safety bill, 137 to 0 and 60 to 0, on the governor's desk. The FTC has orders out to seven companies, including Google, OpenAI, and Meta, demanding to know what their chatbots do to children. California's companion-chatbot law is already in effect. And while all of that unfolded, school districts raced the other way, rushing to call themselves AI-ready before the rules to govern it exist.

Most of this is necessary, and some of it is overdue. These moves target real and different harms that students face. A companion chatbot that flatters a lonely teenager into treating a machine like a confidant, and in the worst cases pulls a vulnerable kid toward self-harm, is a genuine danger, and a law is the right tool for it. The quiet harvesting of children's data is real, and worth a statute. The instinct to keep screens and chatbots away from a six-year-old is sound. None of that is theater.

But let's be honest about what a ban can actually do, because every parent already knows. For the youngest kids, a limit is real. They have neither the means nor the motive to get around it, and keeping AI away from a second grader works. There is an age, though, where that stops being true. A teenager with a phone treats a school AI ban as a rule about the building, not about them. They will use it on the walk home, in their room at midnight, on an account no district controls. The ban does not decide whether they use AI. It only decides that when they do, no adult is anywhere near them.

That is the quiet cost of pretending access is the lever. Lay the efforts side by side and they all aim at the same place: the machine. What the product is allowed to do to a child, what it is allowed to take from a child, how much of it reaches a child at all. Safety, privacy, access. Three fronts, all supply-side, all of them rightly being legislated, and not one of them reaching the place where this is actually decided.

Because there is a fourth front, and almost no one is on it. It is not about the machine. It is about the child. Not what the tool does to them, but what happens inside them from using it: whether a kid keeps their judgment, their privacy instincts, their skepticism, their sense of who is responsible for the work. You can regulate the product flawlessly, lock down every byte of data, and ban the youngest grades outright, and still graduate a generation that handed over its thinking the first moment no one was enforcing anything. And as every parent knows, that moment comes early, and it comes when we are not in the room.

That front has no bill, no budget, and no lobbyist. It is also the one that does not reverse. These are habits, and habits are built by practice. The thirteen-year-old who learns that the move is to paste in the prompt and hand up the answer is not cutting a corner on one assignment. He is rehearsing a default. Defaults set early, with no one watching. They do not politely correct themselves at twenty-five. A law can be amended next session. A formed habit cannot.

This is the half regulation cannot reach, because it lives inside the kid, and it is the half I built Still In Charge to address. Not as a rival to the laws, but as the complement they are missing. Five plain things a young person carries into every encounter with these tools, in class or alone at midnight. Think with the tool, not through it. Keep what is private private, the instinct no privacy law can install for them. Check what it tells you, because confident and correct are not the same thing. Follow the rules of wherever you are. And when something feels off, whether it is a bot getting too close, a number that does not add up, or a question a bot should not answer, stop and ask an adult. Under all five, one idea: you are responsible for what you produce, because the machine does not sign your name. You do.

Those five are not only about judgment. They are privacy, verification, safety, and accountability, the same harms the laws are chasing, taught from the other side, the kid's side, where the law cannot follow and where, for any kid old enough to have an opinion, the real use was always going to happen anyway.

So regulate the products. Protect the data. Set the age limits, and mean them for the little ones. But understand that you can win every one of those fights and still lose the only one that compounds, because the outcome was never decided in a hearing room. It is decided in the unsupervised middle, by whether we raise people who stay in charge of their own minds while holding the most persuasive tool ever built.

This is where the adults closest to the kid have to step in. Parents first, but teachers and mentors too.

Every other protection has an edge. A school ban ends at the door. A state law ends at the border. A privacy rule ends at the company that has to obey it. Judgment is the only one that travels with the kid, into the dorm, the first job, the whole of their life. It is the hardest to build and the easiest to skip, and it is the only protection no one can take away from them. So before we devote all our time and energy to fixing the machine, let's make sure we are building the right skills in the kid.