AI Is Emptying the University of Learning
The script flipped fast. Panic about plagiarism turned into pep rallies for "AI-ready classrooms." Students now draft essays with a chatbot. Professors crank out lectures and grading with the same tool. Degrees start to mean less, while vendors take a bigger cut.
This isn't innovation. It's outsourcing the work of thinking.
The Pivot: Panic to Partnership
One public system spent millions on campus-branded chatbots while cutting programs, staff, and core services. Administrators justified it as "fiscal sustainability" and "student success." The message was clear: we can't fund teachers, but we can fund tools.
That trade doesn't improve learning. It just lowers the cost of pretending.
Tools vs. Technologies
There's a difference. Tools help you do a task. Technologies reshape the task-and then the culture around it. When the metric becomes speed and output, the classroom turns into logistics: generate, submit, grade, repeat.
Thinking turns into prompting. Presence turns into production. Curiosity gets priced out.
The Cheating-AI Feedback Loop
Students automate assignments. Schools buy detectors with high false positives. Then they buy more AI to block the first AI. It's an ouroboros built on budgets and fear.
Some companies now market cheating as a feature. Others quietly normalize it in the name of "efficiency." Meanwhile, a professor bans AI in the syllabus and uses it to make lecture slides. Students notice. Trust erodes.
From Bullshit Jobs to Bullshit Degrees
We're drifting toward a credential economy no one believes in. Students pay for work they didn't do. Faculty grade work they suspect isn't real. Administrators cheer "productivity gains" while learning declines.
Employers receive degrees that signal less and less about actual skill.
Technopoly on Campus
Once the institution treats optimization as the highest good, pedagogy gets replaced by process. Course design becomes prompt templates. Assessment becomes dashboards. "AI literacy" often means "make AI do it."
That isn't education. That's managed output.
Who Pays the Hidden Costs
Behind the glossy demos are real people and real resources. Content moderation has been outsourced to low-paid workers who sift through trauma so the model looks "safe." Training and inference pull huge amounts of electricity and water, often in regions already stressed.
Read the reporting and decide if this is what public education should underwrite: TIME on outsourced AI moderation. Also see the scholarship on how systems embed bias: Algorithms of Oppression (NYU Press).
Students Aren't Buying It
Working-class and first-gen students see the con. They're being asked to pay premium prices for an automated product-one that dilutes mentorship and hollows out the very relationships they enrolled for.
Many are pushing back: "Why would I take on debt if my learning is outsourced?" Fair question.
The Cognitive Debt Problem
Handing off writing and analysis feels efficient in the moment. Over time, it trains the brain to avoid strain. The result is polished output with shallow thought. Remove the tool and performance drops.
Thinking-like strength-grows under resistance. Remove the weight and the muscles atrophy.
What Educators Can Do This Semester
- Set a clear AI policy per assignment. Spell out permitted uses (e.g., idea jogging, outlining) and banned uses (drafting, citations, final language). Require a short "AI use note" on submissions.
- Make thinking visible. Use versioned drafts, process notes, and in-class writing that feeds final work. Grade the process, not just the product.
- Add oral defense moments. Five-minute viva checks expose ghostwriting fast and build real understanding.
- Localize assessments. Tie prompts to lived context, class-only sources, data you provide, or fieldwork. Generic prompts invite generic output.
- Rotate formats. Studio critiques, whiteboard problems, annotated reading notes, and micro-presentations cut through synthetically smooth prose.
- Use AI for critique, not creation. Let students compare AI drafts to their own, mark hallucinations, and fix logic. That builds discernment.
- Require citations that AI can't fake. Page numbers, timestamps, field interviews, or lab notes.
- Protect equity. Don't turn course success into a "who prompts best" contest. Prompt skill ≠ learning.
Assignment Patterns That Resist Auto-Complete
- Case + Countercase: Students analyze a real case you provide, then build and defend a counterfactual.
- Source Triangulation: Synthesize three conflicting sources (one from class). Grade the reconciliation, not the summary.
- Public Annotation: Students annotate a reading in a shared doc pre-class; class time resolves disagreements.
- Make-It-Work Brief: Give messy constraints (budget, audience, medium). Require a one-page rationale with trade-offs.
Academic Integrity Without Surveillance Theater
- Don't rely on detectors. High false positives hurt multilingual and marginalized students.
- Ask for receipts. Process logs, drafts, margin notes, and references beat black-box scores.
- Teach the "why." Grade-linked visas, aid, and jobs create pressure. Address incentives; don't just police behavior.
For Writers: Keep Your Voice
- No AI first drafts. Draft by hand or in a distraction-free app. Use AI later to surface blind spots.
- Voice ledger. Maintain a personal style sheet: phrases you overuse, metaphors you like, topics you avoid. Protect originality.
- Source discipline. Fact-check every "confident" claim. If AI suggests it, verify or cut it.
- Constraint sprints. Time-box short drafts with hard constraints (e.g., 200 words, one argument, one story). Constraints grow craft.
Governance That Puts Learning First
- Faculty-first adoption. No major AI contracts without faculty review, pedagogical pilots, and public reporting.
- Data transparency. Where is student data stored? Who gets access? What labor produced the model?
- Budget honesty. If you can fund licenses, you can fund tutoring, writing centers, and smaller sections.
- Union and senate collaboration. Put AI policy into contracts and curriculum governance, not memos.
The Line Worth Holding
Education is the slow work of building judgment. If we trade that for frictionless output, we get degrees without competence and classrooms without purpose.
Use AI where it supports thinking. Refuse it where it replaces it.
Resources
- Reporting on hidden AI labor and safety work (TIME)
- Algorithms of Oppression by Safiya Noble (NYU Press)
If you need a clear view of AI tools to set smart boundaries and policies for your team, scan this concise index: Popular AI tools (overview). Use it to choose where AI helps-and where it has no place.
Your membership also unlocks:
