AI Killed the Old College Model. Good.
For seventy years, a well-made college essay was proof of the thinking behind it—you couldn't produce the product without doing the work. GenAI severed the two. That's a loss worth having.
Last spring, a professor friend told me she’d stopped assigning whole books. Not because she wanted to, but because her students couldn’t finish them. The same week, her university circulated its third AI policy in two years. Ban it? Embrace it? Detect it? Nobody knew. Nobody knows.
I know this firsthand. In my own doctoral program, I discovered that no one had ever taught me, or most of my classmates, how to read a research paper. We were expected to already know.
Colleges are struggling to figure out what to do with the next generation of students, and the panic has a common root: the implicit contract of higher education has broken. The contract was straightforward. We transfer information, students demonstrate they absorbed it through essays and exams, and everyone agrees to call that an education. AI severs both halves at once. The essays and exams were only ever proxies for learning, and those proxies can now be gamed in seconds. And the thing they stood in for—the processing of information itself—AI now does better, faster, and for free. So the deal fails from both ends: we can no longer tell who learned, and it no longer matters whether they did. If that's what college is for, college is over.
But that was never what college was for. It was for something harder to name and harder to teach: the ability to read a claim and evaluate it. To ask what the evidence is, whether the reasoning logically connects the evidence to the claim, what follows if the claim is true, and who wants you to believe it. That skill has a delivery mechanism, and it isn’t the lecture. It’s deep, critical reading followed by discussion or discourse.
The symmetry is uncomfortable. The technology that makes reading feel optional is the same technology that makes critical readers indispensable. AI produces fluent, confident, sometimes-wrong text at infinite scale. Every hour of your life now involves judging machine-generated claims. A society that can’t read deeply is defenseless.
So this publication makes one argument, from different angles, with a method attached: the more we rely on AI, the more we need people who know how to read and think critically. And that skill can and should be taught.
The method is called CERIC, and it’s five questions a deep reader asks of any text: What’s the Claim? What’s the Evidence? Does the Reasoning hold? What are the Implications? What’s the Context? Five questions, endlessly deep, and entirely AI-proofAI can answer them. The point is that you can. Reasoned judgment, unlike information, doesn’t transfer. It has to be built, one text at a time.
That’s what we’ll do here:
Explore what college is for now
Walk through the CERIC method that you can use on a syllabus, with your students, or alone with a difficult text.
Practice deep and critical readings of texts that deserve our attention—including, sometimes, the output of the machines themselves.
The old college is dead. The case for deep readers and critical thinkers has never been stronger. Welcome to The Deep Reader.
Read Deeply,
Dr. Genevive
If you teach, parent, or read — subscribe. Next week: The Students Who Can’t Finish a Book.
A question for the comments: if the old model is dead, what’s the one thing college still does that nothing else can, or is even that gone?