I see a lot of people conflating "AI literacy" with "thinking with AI" and those are not remotely the same thing. Kids should understand AI. They should know what it is, what it can do, where it fails, how to question it and how not to be fooled by it. But using AI as a substitute for reasoning is something else entirely. And I think we are being incredibly naive about what children are actually going to do with it. I have 9 kids. They all have very different personalities. But they all share one thing. If there is an easier way to do something they will take it. Of course they will. They are kids. Adults do this too. The difference is that most adults grew up before there was a machine sitting next to them that could instantly remove the hardest part of almost any intellectual task. That matters because the hard part is not some annoying obstacle standing in the way of learning. The hard part is the thinking. And it's the most important. Trying to understand something when you don’t get it. Sitting with confusion. Making a bad argument and realizing it sucks. Searching for the right word. Comparing one idea to another. Changing your mind. Starting over. Finishing something when you are sick of working on it. This is how people learn to think. We can give these things big names like metacognition, analogical reasoning, epistemic justification and abstraction. But the ideas are simple. Can you examine your own thinking? Can you connect one idea to another? Can you explain why you believe something and not just the belief itself? Can you keep working when you don’t know the answer yet or realize the answer you've been given may be factually correct but filled with presuppositions? These are learned abilities and you learn them by DOING them. This is what a lot of adults are missing. If you grew up pre-AI then you already spent twenty or thirty or forty years developing a mind without it. You had no choice. You wrote the paper yourself. You stared at the blank page. You solved the problem. You read things you didn’t understand. You got bored. You made bad arguments. You learned analogies and idioms. You worked toward goals. You failed. You slowly built judgment. Then AI showed up after all of that development had already happened. Now those same adults look at children and say "AI is just another tool. They’ll use it the same way I do." No they won’t. That makes no sense. You cannot take a fully developed adult mind and add AI to it then assume a ten year old can start with AI and somehow arrive at the same intellectual destination or modus of use. The order matters. A lot. And this is the part people keep skipping. A calculator is useful once you understand numbers. Spellcheck is useful once you understand language. AI goes way further than these tools. It generates the reasoning. The argument. The explanation. The analogy. The thesis. The counterargument. The conclusion. It is not just helping with the execution of thought. It can replace the formation of thought itself. And kids will use it for that everytime. That should scare the hell out of us more than it does. Thomas Kuhn wrote about paradigm shifts. Changes so big that they don’t just give us new answers. They change the framework we use to understand the questions in the first place. AI is that kind of change. Yet most of the conversation around children is still about productivity. Can they learn faster? Write faster? Research faster? "Create" faster? That completely misses the point. For children this is a developmental and socialization issue. The question is not whether AI can help a child complete an assignment and they "cheated". Of course it can. The question is what happens to a child who doesn't have to do the difficult mental work to arrive at a newly unthought of conclusion. What kind of person do you get after ten or fifteen years of always having an immediate way around confusion and boredom and difficulty and failure? I think you get people who look incredibly capable right up until the moment something becomes difficult or ambiguous or unconventional. David Goggins has a phrase he used about the pre-SEAL training program. He said all it produces is "Faster, stronger, bigger quitters". That is exactly what worries me. People who can produce more. Move faster. Operate more tools. Assemble more information. Create polished work. But they have almost no tolerance for being stuck and worse cannot find new ways of thinking. Every decision will be built on a prior pattern. That is what AI is. And intelligence is not the only issue. What about fortitude? The ability to stay with something when quitting feels better. The ability to hold an unpopular belief because you actually reasoned your way into it. The ability to zig when everyone else zags. To play the dissonant note. To follow an idea that looks stupid before it starts looking interesting. AI does not teach you any of that. Difficulty does. Failure does. Boredom does. Having to figure shit out yourself does. There is research around brain regions like the anterior mid-cingulate cortex and their connection to effort and persistence and doing things you would rather stop doing. But you don’t even need the neuroscience to understand the point. You do not build fortitude by removing every occasion that requires fortitude. If a child has an instant escape hatch from intellectual difficulty they are going to use it. A lot. This is also why I don’t buy the idea that kids are mainly going to use AI to "create" or "study better". Some will. Most will use it to avoid. Avoid thinking. Avoid writing. Avoid being wrong. Avoid struggling. Avoid not knowing something. Avoid the boring middle part of actually getting good at something. That is human nature and anyone who has raised kids into adolescence knows exactly what I mean. What I find genuinely bleak is the idea of raising a generation that can generate an answer to almost anything but cannot explain why they believe what they believe. People who are insanely fluent at prompting machines but strangely inarticulate when you ask them "What is wrong with this way of thinking?" People who can prompt a position and refine a position and defend a position that an agent helped construct but never really learned how to arrive at one themselves and then when asked to, they turn to the same model to do so. AI literacy matters. Kids should understand the technology that is going to shape their lives. But before children learn to think with AI (because that is what we are all doing) they need to learn how to think without it. Those are not the same thing and if we get that order wrong we are going to regret it.