Raising the Bar: Why AI Doesn't Erode Rigor, It Redefines It

Over the past month, a burst of essays has warned that artificial intelligence is hollowing out education. In July alone, Meghan O’Rourke argued in The New York Times that AI robs students of the slow, uncertain struggle with language that produces discovery. David Brooks cautioned that it seduces us with “excellence without effort.” And a student writer, John Cracraft, described campuses where curiosity itself seems to be evaporating. Meanwhile, a new MIT study seemed to confirm the worry, claiming students who relied on AI showed weaker memory and brain connectivity.

The anxiety is real. But the diagnosis is wrong.

AI does not eliminate rigor or creativity. It changes where they live. What was once “remarkable” becomes the new baseline. Every major intellectual tool in history has worked this way. Writing, the printing press, photography, the internet — all lowered one threshold while raising another. The pattern is not erosion, but inflation: creativity inflation.

The old bar and the new floor

Consider the prompts used in the MIT “cognitive debt” study. Students were asked to write on questions like: Does true loyalty require unconditional support? or Do works of art have the power to change lives? Unsurprisingly, essays produced with AI were smoother but left students with less recall. But let’s be honest: these are SAT-style prompts designed to test baseline fluency, and AI now dispatches them with ease.

The problem is not that AI makes students shallow. The problem is that the bar was already too low. A generic five-paragraph essay is no longer evidence of rigor; it is evidence you can prompt a chatbot.

If you asked Meghan O’Rourke or Douglas Hofstadter to write on the same topic under the same rubric, their essays would not be “competent.” They would be remarkable. Every sentence weighed. Every transition deliberate. Every word chosen to carry meaning. That is the new bar.

But even that won’t hold forever. Because if O’Rourke or Hofstadter themselves begin to use AI as writing partners, the expectation for them rises again. It won’t be enough to produce beautiful essays. The new bar will be:

• Density of meaning — prose where every line resonates on multiple levels: philosophical, emotional, rhetorical.
• Structural complexity — recursive or layered architectures of argument that an AI can scaffold but only human craft can orchestrate.
• Cross-domain synthesis — fusing poetry with neuroscience, philosophy with personal narrative, history with computation. The machine can surface connections across fields instantly, but the human must decide what holds together.
• Curation at scale — not just drafting once, but iterating across dozens of AI variants and exercising judgment to select what is extraordinary, not merely competent.
• Meta-transparency — writing that reveals its own making, showing where the AI was scaffolding and where the human voice transformed it.

I see this already in my own world: software development. Before AI coding assistants, a good developer might write at about 30 words per minute, carefully crafting logic line by line. That was considered excellent. Today, if my team wrote code at that pace, they wouldn’t be employed. The baseline has changed. AI handles boilerplate instantly; the new bar is architectural judgment, orchestration, and debugging at scale. The job isn’t gone, but the expectations are far higher.

This is the pattern we have seen for millennia. Each time the floor rises, excellence migrates upward. What AI makes “remarkable” today will become tomorrow’s baseline. The masters will not be dethroned. They will be forced into a higher register of artistry, one that no baseline draft, human or machine, could ever achieve.

Where the critics go wrong

1. Loss of process? O’Rourke worries that AI shortcuts the slow wrestling with language that makes writing formative. But syntax and revision aren’t obsolete — they’re more essential. Because AI drafts are competent but bland, students must cut clichés, reshape sentences, and inject rhythm to produce something that stands out. Revision doesn’t vanish; it intensifies.

2. Cognitive debt? Brooks points to brain scans showing less activity in students who let AI do the work. But this is obvious: if you outsource the whole assignment, of course you recall less. The fairer question is: what happens when students are required to revise, critique, and personalize the AI draft? That work still demands the same, if not greater, cognitive effort.

3. Erosion of curiosity? Cracraft sees classmates asking only, “How do I get this done fastest?” But that malaise predates AI. It is the product of incentives — grades, résumés, compliance. AI simply exposes the hollowness of low-level assignments. Curiosity doesn’t die when AI enters the room; it dies when we reward students for stopping at the baseline.

4. Derivative output? O’Rourke likens AI to “intellectual Soylent Green” — a paste made from the bodies of writers. She’s right that AI is derivative. But so is all culture. Every poet builds on predecessors. Jazz sampled blues; hip-hop sampled jazz. The question is not whether work is derivative but whether it adds something new. That’s the bar humans must clear.

5. Flattening of excellence? All three writers fear that standards will decline. Yet history shows the opposite. After the calculator, arithmetic drills lost value but higher-order math soared. After Google, memorizing dates lost value but synthesis and sourcing rose in importance. AI makes mediocrity cheap. It makes excellence more demanding.

Where the new bar lies

So where is rigor now?

• Syntax and style. Fluency is baseline. The bar is in rhythm, diction, surprise — sentence-by-sentence craft.
• Revision and refinement. A single draft is worthless. Students must show multiple iterations, sharper cuts, tighter flow.
• Original insight. AI produces consensus. The bar is contradiction, paradox, risk. These are the moves machines will not make.
• Personalization. AI cannot import your lived experience. The bar is weaving universal arguments with personal detail.
• Meta-cognition. Students must defend their changes, explain why they diverged from AI, and reflect on process.

That is not less demanding. It is more demanding. It engages as much “brain activity” as ever, only at a higher level.

The future of excellence

The truth is simple: AI does not end rigor. It resets the floor. The question facing educators, leaders, and students is not whether AI makes us weaker. It is where to set the new ceiling.

The five-paragraph essay was once the marker of college readiness. Soon it will be the marker of middle school. Yesterday’s Pulitzer-level essay will become tomorrow’s college standard. Yesterday’s Oscar-worthy screenplay will become the baseline for television pilots. That is creativity inflation.

Excellence has never been static. Each time a tool lowers the cost of mediocrity, we demand more from those who aspire to greatness.

Takeaway

The critics are right about one thing: the ground has shifted. What counted as hard yesterday is trivial today. But the conclusion they draw is wrong. Writing isn’t dying. Curiosity isn’t dying. Rigor isn’t dying.

The bar has moved.

And the challenge for students, professionals, and even masters is to rise to meet it.

---

A confession: I’m not a natural writer. This essay took me more than ten rounds of drafting and reshaping with AI to get right. Which is, in a way, the point. AI didn’t save me from the work — it forced me into it. Every iteration raised the bar a little higher until I finally had something worth sharing. And perhaps that’s the uncomfortable truth: what feels like “high effort” for me today may be too low a bar for Meghan O’Rourke’s future students. In the AI age, even struggle itself gets inflated. The floor keeps rising, and so must the ceiling.