This spring, Florida joined a growing wave of individual states convening task forces to write guidelines for AI use in schools. It’s a necessary step but a slow one: formal guidance will take a year or two to reach the classroom. Meanwhile, at home, something is already changing. Homework that used to take forty minutes now takes eight. Your kid’s essay is perfect. But have they actually learnt anything from it?
Your Kid’s Homework Looks Great, Or Does It?
AI produces confident, well-structured answers on demand. But fluency isn’t evidence of understanding. When a student hands a hard question to an AI tool, the tool does the part of learning that builds the brain: sitting with a problem that doesn’t make sense and working through it until it does. Pass difficult questions to AI enough times, and gaps in learning become visible, at least according to emerging studies.
In a study of 1,000 students, students using an AI math tutor outscored a control group by 48% while the tool was in front of them, but scored 17% worse than the control group once it was taken away. The tool made their work look better, but the students were not learning from it.
Researchers have a name for what’s happening: cognitive offloading, the habit of handing a mental task to an external tool rather than doing the thinking yourself. Offloading isn’t new, we’ve always used calculators and calendars, but AI offloads on a much larger scale than other tools could. It’s not just calculations or dates; it’s the framing of an argument, a comprehensive analysis, or working out a complicated math problem. The more of that judgment a student outsources, the less practice their brain gets forming it independently.
The thing to note is that kids don’t notice this happening to them. Kids may feel that they are learning something even while using AI to generate answers. Fluent output feels like understanding, whether or not they produced it. Parents often can’t catch it either, especially in higher-grade material they haven’t touched since their own school days.
Is AI Your Kid’s Crutch or Coach?
With AI’s help, it may look like your child is finally producing sophisticated ,thoughtful responses. But polished work does not necessarily reflect cognitive growth. Is your child still doing the thinking, or is AI doing the heavy lifting?
That’s especially true for kids who are already reluctant to sit with challenging material. Because they find it difficult, they avoid tasks that demand sustained effort, or they don’t fully trust their own ability to work through difficulty.
Why Productive Struggle Is the Key
When a student works at the edge of what they can currently do, the brain strengthens the neural connections that make the next hard thing easier. Remove that friction too early by utilizing AI means that the brain never gets the training.
Before students can use AI effectively, they need the underlying cognitive skills to use it well: attention span, working memory, reasoning, hypothesis-forming, critical thinking. Tara Bonner, Arrowsmith’s Chief Neuroeducation Officer, points out that most AI guidance for schools covers what kids should know about the technology, prompting, citation, academic integrity, but almost nothing about whether a student’s brain is actually ready to use it well. Without that readiness, over-reliance on AI can create what’s increasingly called “cognitive debt”: a gradual erosion of the brain’s ability to hold onto, retrieve, and connect what it’s supposed to be learning.
The encouraging news is that this is trainable. Decades of neuroscience show the brain can change. By strengthening the specific cognitive functions underneath learning, students improve their capacity to learn, understand, and perform. That’s the premise of the Arrowsmith Program: a cognitive training program used by schools, educators, and parents to build their students’ cognitive capacity across every subject that depends on it.
What to Watch Out for This Term
You don’t need to grade the homework yourself. At a certain point, most parents are quite happy to leave the calculus, chemistry, or new approach to long division, to the teachers. But you can watch for a few things. Can your child explain, out loud, why an answer is right? Can they point to something in the AI’s response that’s actually wrong? Can they redo a similar problem with the app closed?
One simple house rule is worth adopting: attempt the question first, and use AI second. Treat AI as something that checks work, not something that does the work.
If what you’re seeing looks less like a motivation problem and more like a real gap, if your child avoids anything hard because effort hasn’t paid off for them before, that’s a signal of weak underlying cognitive functions. This is where an approach like the Arrowsmith Program can come in to support.
The Real Question This Semester
Florida’s task force will eventually produce guidance, but it will arrive well after this semester is over.
Your child’s habits are forming now. You don’t need your school’s framework to ask the question: Did AI help my child get better at thinking? Or did it just help them get better at completing homework?
AI is only going to get more capable. The one thing you actually get a say in is whether your child’s’s mind is getting more capable right alongside it.
