Your daughter is stuck on a word problem. She opens an AI tool, types the question, and two seconds later has an answer. She moves on. You feel relieved — one less thing to re-learn tonight.
But here’s the question worth pausing on: did she just learn something, or did she just get past something?
That’s not a rhetorical trick. It’s the whole test we use at Auxesis for any tool that touches a kid’s education: does this augment thinking, or replace it? Does it deepen learning, or shortcut it? Everything below comes back to those two questions.
AI is already in most homeschools — the question is which kind
Homeschooling has grown fast. Roughly 3.4 million children were homeschooled in the US for the 2025–2026 school year, and the growth rate nearly tripled compared to before the pandemic (Johns Hopkins Institute for Education Policy). Alongside that growth, AI adoption has followed — recent industry estimates put AI use among homeschooling families at around 40%.
That’s a lot of families making a judgment call with very little guidance. Most of what’s out there is either a best AI tools listicle or a general AI safety for kids article that isn’t written with homeschooling in mind. So here’s a framework built specifically for this decision: five red flags to check before an AI tool gets anywhere near your child’s curriculum.
It gives the answer before it asks a question.
This is the single biggest tell. A tool built to teach will slow a student down — ask what they’ve tried, offer a hint, wait. A tool built to please will just solve it. Even market leaders struggle here: reporting on Khan Academy’s Khanmigo found that if a student simply repeats I don’t know a few times, it can be prompted into handing over the answer outright — something a human tutor is far less likely to do. If a tool never makes your child sit with not-knowing for a moment, it’s optimizing for a finished worksheet, not a working brain.
Red flag two: there’s no visible teaching process.
Ask yourself: could you explain, in one sentence, how this tool teaches? If the honest answer is it just answers whatever you type, that’s not pedagogy, that’s autocomplete. Look for a tool that shows its method — the Concrete-Pictorial-Abstract progression in math, guided questioning in reading, a visible here’s why we’re doing it this way. No visible method usually means no method.
Red flag three: it collects more than it needs to.
Some AI learning tools quietly gather names, voice recordings, location data, or browsing behavior — sometimes to improve the product, sometimes to train other models, not always with clear consent. Before letting a tool near your child, check whether it says plainly what it collects and why. If that information is buried or missing, treat that as the answer.
Red flag four: there’s no supervision or moderation layer.
A tool built for kids should say so clearly and back it up with real controls — content filtering, session limits, a way for you to see what was asked and answered. If a product markets itself to students or learners in general rather than children specifically, assume it wasn’t built with your child’s age in mind.
Red flag five: it’s confident even when it’s wrong.
AI tools rarely say I’m not sure. They answer everything with the same even tone, whether the material is a well-established math fact or a shaky guess. That’s a real risk for a subject like history or science, where a confidently wrong answer can lodge itself as fact. A tool worth trusting will hedge, cite, or flag uncertainty — not just answer.
The stakes are real, and they’re not just about grades
A growing body of research backs up why this matters. The OECD’s Digital Education Outlook 2026 documented a measurable decline in critical thinking tied to what researchers call cognitive offloading — the more students lean on AI to do the thinking for them, the less practice they get at doing it themselves. Khan Academy’s own data tells a related story: despite Khanmigo logging over 108 million interactions since 2023, only about 15% of students with access to it use it regularly — a low-engagement number significant enough that Khan Academy is rolling out a full redesign in summer 2026.
None of this means AI has no place in homeschooling. It means the kind of AI matters more than the fact of AI. A tool that makes your child wait, think, and explain their reasoning is doing something fundamentally different from one that hands over the answer — even if both look identical on the surface.
Where do you draw the line?
We’re curious how other homeschool families are actually navigating this day to day — not in theory, in practice. We started a thread in the Auxesis community asking exactly that: how are you using AI with your kids? Come add your experience — what’s worked, what’s made you uneasy, what you’ve had to walk back. We’ll be reading and replying.
Leave a Reply