The AI Academic Paradox: When Students Forget to Read Their Own Essays

Itis 11:30 PM, I am grading papers, coffee cold, eyes burning. Suddenly, I see a submission that ends with “written with Writesonic.” Not buried in some acknowledgments section where you might miss it. Not even trying to hide it. Just sitting there like the kid had written “Jimmy Thompson (my mom helped)” at the top of a middle school book report. That’s when I realized we’re all living in some weird alternate reality where half the academic world is using AI and the other half is pretending it doesn’t exist.

5/8/20243 min read

A spiral-bound notebook lies on a white surface, accompanied by two green pencils. Several crumpled pieces of paper are scattered around, suggesting a creative or brainstorming environment.

Everyone’s Got Their Head in the Sand

I have never seen anything mess with higher education like this AI stuff has. It’s everywhere and nowhere at the same time.

Publishers are frantically writing new policies. Scientific journals are adding disclaimer requirements faster than you can say “ChatGPT.” Meanwhile, walk into any faculty meeting and listen to the conversations:

“Oh, I would never use artificial intelligence for my research…” (pause) “…but hypothetically, if someone were struggling with their discussion section…”

It’s like watching your parents discover Netflix while insisting they still prefer renting DVDs from Blockbuster.

The whole thing would be funny if it wasn’t so completely bonkers.

Here’s the Thing That Keeps Me Up at Night

AI has actually done something incredible for science writing. And I mean that.

Take my colleague, Maria, a brilliant doctor who does groundbreaking work on infectious disease patterns. However, English is her third language, and academic writing in English can be brutal. Now she can take her complex ideas and get help making them sing on the page. The science is all hers, but the clarity? That’s gotten a massive upgrade.

Or think about researchers who are dyslexic, or who struggle with the weird formal conventions of academic prose. Suddenly, they can focus on what they’re actually good at instead of getting bogged down in whether they need a semicolon or a comma.

This should be amazing news, right?

So why do I feel like we’re all stumbling toward some kind of disaster?

My Brilliant Plan (That Partially Backfired)

Last semester I decided to tackle this head-on instead of pretending it wasn’t happening. I redesigned my whole scientific writing course around this two-part assignment that I was pretty proud of.

Part one: Write everything longhand first. Actual pen, actual paper. No computers, no phones, just you and your thoughts. Then type it up exactly as written.

Part two: Now use whatever AI tools you want to polish that original work into something publication-ready.

The idea was to keep the human thinking while embracing the technological help. Most students totally got it — they improved their arguments, caught errors, learned about revision. It was working exactly like I’d hoped.

But then…

The Ones Who Completely Missed the Point

Some students just… didn’t do part one at all. They went straight to feeding prompts into various AI tools, let the machines generate entire papers, and then submitted whatever came out.

The Writesonic thing was just the most obvious example. I found papers with internal contradictions, citations to studies that don’t exist, methodology sections that made no sense with the supposed results. One paper concluded that “further research is needed” about something the introduction claimed had already been definitively proven.

They hadn’t just cheated. They’d cheated so lazily they forgot to proofread.

But here’s what really got to me — they were essentially paying tuition to avoid learning anything. It’s like hiring a personal trainer and then taking a nap during every session.

The Grown-ups Are Just as Bad

Meanwhile, in the “professional” world, researchers are quietly using AI for everything — grant applications, literature reviews, first drafts of papers. They’ll admit it after a few drinks at conferences but never put it in writing.

This weird collective denial isn’t helping anyone. When we can’t talk honestly about how these tools are being used, we can’t develop good practices or teach students to use them responsibly.

We’re all just making it up as we go along and pretending we know what we’re doing.

What Actually Needs to Happen (In My Humble Opinion)

I am not anti-AI. That would be stupid and pointless, like being anti-Internet in 1995. However, we need to act together to solve the problem.

First: Stop pretending it doesn’t exist. Acknowledge how it’s being used, create some actual guidelines, and have honest, open conversations about the best possible practices.

Second: Figure out the boundaries. AI for grammar and clarity? Great. AI for generating your hypotheses and doing your thinking? Not so great.

Third: For the love of all that’s holy, read your own work before you submit it. If you can’t explain what you’ve written, you probably shouldn’t be putting your name on it.

The Real Problem Isn’t the Technology!

The student with the Writesonic signature taught me something important. The problem isn’t that AI exists or even that students are using it. The problem is when people become so disconnected from their own work that they don’t even bother to read what they’re submitting.

That’s not a technology problem. That’s an engagement problem.

If we’re not willing to read our own papers, why should anyone else?

Have you run into this stuff in your work? I’m curious whether other people are experiencing the same kinds of issues, or if I'm just unlucky with my students. Leave a comment — and yeah, I promise I’ll actually read them all myself.

Connect

Explore science, creativity, and shared passions here.

baruh@polisstory.space