A Reddit thread went viral last week. Teachers, hundreds of them, drawing a direct line between today’s AI panic in classrooms and the Wikipedia panic of the mid-2000s.
"They told us not to use Wikipedia because it wasn’t reliable. Now they’re telling us not to use AI because it isn’t reliable."
The thread has been climbing for days. The vibe is exhaustion — and a growing sense of déjà vu.
Here’s the thing: they’re right. And if you know the history, you also know what happens next.
We’ve Done This Before
In 2007, a New Jersey school district blocked Wikipedia on every school computer. Librarians posted "Just Say No to Wikipedia" signs. The concern was real and widespread: students were using Wikipedia as a primary source, copying text directly, skipping the actual research process.
Ars Technica covered the bans at the time, noting the irony — students had access to Wikipedia from home computers anyway, so the ban was less about accuracy and more about a fundamental discomfort with a new kind of information tool.
Before that, it was calculators. In April 1986, math teachers picketed outside a national math-teaching conference in Washington, D.C., protesting a new policy promoting calculator use in classrooms, even for young children. The fear: children would forget how to do arithmetic. Schools restricted them. Parents worried. It became its own small culture war.
Sound familiar?
What the History Shows
Every time a new information tool has entered the classroom, the same pattern has repeated:
- Panic. The tool is dangerous. It will replace real thinking. It will corrupt the learning process.
- Restriction. Schools ban it. Teachers warn against it. Parents are counseled to limit it.
- Integration. Eventually, the tool gets absorbed — selectively, thoughtfully — into good pedagogical practice.
- Retrospective. A decade later, people wonder what the fuss was about.
Researcher Amy Orben’s paper "The Sisyphean Cycle of Technology Panics" (2020) documents this exact pattern across decades of new-technology fears — and her conclusion is that the panic is predictable, the resolution is also predictable, and the skill that actually matters throughout is the same one it always has been.
Judgment.
What Khanmigo’s Numbers Actually Tell Us
Khan Academy recently disclosed that only 15% of students with access to Khanmigo regularly use it — despite 108 million interactions since launch. Their response is to redesign the product to be more proactive, to guide students through assignments rather than waiting for them to initiate.
This is not the story of a tool that failed. It’s the story of a tool that was built around the wrong assumption: that students would seek out AI help on their own.
Students, it turns out, behave the same way with AI tutors as they do with calculators, Wikipedia, and every tool before them. They use them when the friction is low and the value is obvious. They avoid them when the interaction feels performative or disconnected from what they actually need.
The 85% who aren’t using Khanmigo aren’t proving AI doesn’t work in education. They’re proving that AI-as-answer-machines don’t work in education.
The Wikipedia Lesson, Applied to AI
The question was never "should students use this tool." The question was always: "what kind of thinking does this tool require, and is that the thinking we want to develop?"
Wikipedia didn’t make students worse at research. Students who used Wikipedia thoughtfully — who cross-referenced, who read the citations, who asked why an article was written the way it was — got better at evaluation. Students who treated it as an answer dump got better at dumping answers.
AI is the same, but faster, and with higher stakes.
The tools that will survive this cycle — and the ones worth building practices around — are the ones that require genuine cognitive engagement. Not "give me the answer." But: "walk me through how you’d think about this."
That’s a fundamentally different design. And it’s the only one worth building toward.
What This Means for Your Family
If you’re a parent trying to figure out what to do right now: the noise is temporary. The history is consistent.
The families that will come out of this moment ahead are not the ones who banned AI outright. They’re the ones who started the harder, more valuable conversation: what does it mean to use this tool well?
That’s not a technology question. It’s a thinking question. And it’s one worth sitting with.
Auxesis helps families build that thinking — with AI tools designed to scaffold discovery, not replace it. Explore what that looks like at auxesislab.com.
Leave a Reply