The AI-Psychology Crisis in the Classroom
It's not about cheating. Not really.
When Brown University professor Roberto Serrano saw his students' midterm scores climb to 96 percent — a jump from the usual 65–80 — he didn't panic. He waited. He knew better than to trust a spike in grades. He knew the real test would come later, under supervision. And when it did? The average plummeted to 48.6. Three students scored zero. That's not a fluke. That's a fracture.
This isn't the first time AI has exposed a rot in how we measure learning. But it's the first time the rot has been so loud, so visible, so personal. Serrano didn't just call out students. He called out a system that had quietly surrendered: "We cannot choose to become idiots." That line didn't come from a press release. It came from a man who'd spent twenty years teaching, watching students grow, and now watching them outsource their thinking to machines they didn't understand.
This is where AI in psychology becomes unavoidable. We're not just seeing students cheat. We're seeing their cognitive architecture rewire. The same neural pathways that once built problem-solving stamina now bypass effort entirely. The dopamine hit from a quick, AI-generated solution replaces the slower, richer reward of figuring it out yourself. And here's the cruel twist: the students know this. Eighty-eight percent of Brown's undergrads admit they're afraid AI is making them dumber. They're not lying. They're terrified.
The university didn't respond with bans. They didn't scream about "academic integrity" like a broken record. They looked at the data. They found that 67 percent of grad students use AI weekly. That 85 percent of master's students do. And yet — and here's the kicker — the same students who use it daily also report that it's hurting their ability to think deeply. They're addicted to the crutch, and they hate how weak it's making them.
This isn't a policy problem. It's a psychological one. And if we treat it like a policy problem — more proctors, more software, more rules — we'll lose. Because the real enemy isn't ChatGPT. It's the quiet erosion of effort. The belief that thinking should be effortless. The assumption that intelligence is something you can rent, not build. As the calculator analogy for AI shows, treating AI as a neutral tool ignores the fundamental difference between offloading computation and offloading judgment — and that distinction is exactly what's at stake here.
The Market Correction Is Already Happening
Here's what no one wants to say out loud: the value of a college degree isn't measured by the number of students who pass. It's measured by the number of students who earned it.
Employers aren't stupid. They've seen the grades. They've seen the resumes glowing with "AI-assisted" projects. And they're starting to ask: if this student couldn't write a five-page paper without a bot, what happens when they're on the job? When there's no algorithm to rescue them from a crisis? When the stakes are real?
We're already seeing the market react. Companies like Goldman Sachs and McKinsey quietly started asking for handwritten take-home assessments. Not because they're Luddites. Because they're terrified of hiring someone who can't think without a crutch. A degree from a school that lets AI do the heavy lifting? That's a liability.
Meanwhile, institutions like Brown — the very ones caught in the scandal — are now leading the charge toward rigor. They're not banning AI. They're teaching students how to use it without letting it replace thinking. They're designing exams that can't be outsourced: oral defenses, in-class problem-solving under time pressure, projects that require iterative feedback and personal reflection. They're betting that the long-term value of their degrees will rise precisely because they're harder to fake.
This isn't theory. It's economics. When a product's quality signal is clear — when employers know that a degree from X University means the holder actually learned something — demand increases. And when the signal is muddy? The market punishes you. Hard.
The universities that cling to the illusion of accessibility — the ones that say "AI is just a tool" and then do nothing — will become the community colleges of the 2030s. The ones that demand rigor? They'll be the ones recruiters line up for. The market doesn't care about your intentions. It cares about your outcomes. And right now, outcomes are being rewritten by the students' own cognitive decay — a phenomenon explored in the brain recalibration effect, where digital tools raise our threshold for what feels like "enough" effort, making genuine intellectual work feel disproportionately expensive.
The Illusion of Neutrality in the Age of AI
Let's be honest: most universities are terrified of taking a stand.
They don't want to be seen as "anti-tech." They don't want to alienate students who think AI is their lifeline. They don't want to be the "mean school." So they mumble about "responsible use" and hand out vague guidelines that no one follows.
But here's the truth: neutrality is a myth.
As Kwame Anthony Appiah wrote in The Atlantic, neutrality isn't the absence of values — it's the performance of them. When a university says "we don't take sides," what it's really saying is: "We don't value critical thinking enough to defend it."
The professor who lets students use AI on every assignment isn't being kind. They're being complicit. They're saying: "Your thinking isn't worth the effort." And the students? They hear it. They internalize it. And then they stop trying.
The most dangerous thing about AI in education isn't the tool. It's the surrender. The quiet, collective decision to treat cognitive effort as optional. To treat the mind as something that can be outsourced like a cleaning service.
We're not teaching students to use AI. We're teaching them that their brains are broken. And if you believe your brain is broken, why bother fixing it?
The universities that survive this era won't be the ones with the fanciest AI labs. They'll be the ones that refuse to let their students become idiots — because, as the hidden cost of AI optimization demonstrates, every convenience we accept erodes the friction that makes us flourish.
Credibility Is the Only Currency That Matters
A degree from a top-tier school used to mean something because it was hard to get. Not because the admissions office was picky — because the work was hard. Because you had to wrestle with ideas. Because you had to fail, rewrite, struggle, and finally, finally, understand.
That currency is evaporating.
But here's the beautiful, brutal truth: it can be rebuilt.
It's not about banning ChatGPT. It's about redesigning the entire assessment ecosystem. It's about exams that require live reasoning. About papers that demand multiple drafts with in-person feedback. About capstones that can't be completed without sustained intellectual labor.
Brown's response — the GAITL committee, the focus on AI literacy, the shift toward human-centered evaluation — isn't just policy. It's a declaration. We believe your mind matters. We believe your effort matters. We believe your thinking is worth something.
And that belief? That's what employers will pay for.
The long-term winners in higher education won't be the ones who adapt fastest to AI. They'll be the ones who remind students — relentlessly, consistently, beautifully — that intelligence isn't something you download. It's something you build. Brick by brick. Thought by thought. Effort by effort.
If you want your degree to mean something in 2035? Start demanding more from your professors. Start refusing the shortcut. Start believing your brain is worth the struggle.
Because the market isn't waiting. It's already deciding.