AI Changed the English Classroom. Our Assignments Need to Change Too.

I don’t think the answer to AI is making every student write every essay by hand. I also don’t think the answer is pretending nothing has changed. Because it has.

For a long time, English teachers could assign a literary analysis essay, send students home to write it, and reasonably assume that the paper they turned in represented some version of their own thinking. Maybe they got too much help from a parent. Maybe SparkNotes made an appearance. Maybe they wrote the whole thing at 11:47 p.m. the night before. But generally, if a student submitted a coherent essay, they had to actually produce a coherent essay.

That assumption is gone.

A student can now generate a polished response in seconds. So I don’t think the most useful question is “How do I catch AI?” I think the better question is: “How do I design the writing process so that I can actually see my student’s thinking?”

The Problem Isn’t Just Cheating

AI detection gets most of the attention, but I think there’s a bigger instructional problem underneath it. If an assignment can be completed successfully by pasting the prompt into a chatbot and submitting whatever comes back, the final product is no longer giving us reliable information about what the student actually knows or can do.

That doesn’t mean the essay itself is bad. Plenty of assignments worked perfectly well before generative AI existed. The environment changed.

So instead of throwing out essays—or making students write absolutely everything by hand—I think we need to make the process harder to outsource.

My Formula for Minimizing AI in Essays

This is the system I landed on because I wanted something practical. I did not want to spend my life running essays through AI detectors, interrogating students, or trying to prove where every suspicious sentence came from.

So I built the proof of process directly into the assignment.

Step one: the first draft is a handwritten timed write.

Students write their first draft by hand, in class, under a time limit. It does not need to be perfect. In fact, I don’t want it to be perfect. I want to see what they can actually produce when the ideas and sentences are coming from them.

That handwritten draft gives me a baseline. I can see their argument, their evidence, their organization, their sentence structure, and—most importantly—their voice.

Step two: the second draft is typed in a Google Doc.

Students then take that handwritten draft and type their revised version into Google Docs. Because the second draft is written in Google Docs, I can use the Revision History Chrome extension to replay the writing process and watch the keystrokes if something looks suspicious.

That means I am not relying on an AI detector telling me that a paragraph is “82% likely to be AI.” I can actually look at how the document was created.

Did the student type and revise naturally over time? Did an entire polished paragraph suddenly appear? Did the writing develop in a way that makes sense?

That evidence is infinitely more useful to me than a detector score.

Step three: students highlight every change on the handwritten draft.

This is the piece that makes the whole system manageable.

As students revise, they highlight the places on their original paper draft that they changed in the typed version. Now I don’t have to sit there comparing two entire essays line by line.

I can scan the handwritten draft, see exactly where the student made changes, and then jump to those sections in the Google Doc.

Suddenly, revision is visible.

I can see what they changed, how they changed it, and whether the new version makes sense as an evolution of their own writing.

And if something changes from a rough handwritten sentence to a graduate-level paragraph that sounds absolutely nothing like the rest of the paper? Now I know exactly where to look.

Why This Works Better Than Trying to “Catch” AI

What I like about this system is that it doesn’t make AI the center of the assignment.

The assignment is still about writing.

Students still have to draft, revise, clarify, rethink, and improve. The handwritten timed write gives me authentic baseline writing. The Google Doc gives them the freedom to revise digitally. The highlighted changes make that revision easy for me to track. And the revision history gives me another layer of evidence if I need it.

Most importantly, I’m not punishing every student because some students might misuse AI.

They still get to type. They still get to revise. They still get to use the tools writers actually use.

I just get to see the road between draft one and draft two.

Stop Grading Only the Finished Product

This is the larger shift I think AI is forcing us to make.

A polished final essay used to tell us quite a bit. Now, not necessarily.

So I’m much more interested in everything that happens before the final draft. What did the student write during the timed write? What did they change? What feedback did they respond to? How did their thesis improve? Where did their evidence get stronger? Can they explain why they made a change?

The essay stops being one mysterious document that appears in Google Classroom on Friday night.

Instead, it becomes the final step in a visible trail of thinking.

And honestly, that is better writing instruction anyway.

Make the Work Harder to Outsource, Not Harder to Complete

There’s a difference.

When teachers get frustrated with AI, it’s tempting to lock everything down. No computers. No homework writing. No Google Docs. No outside resources. Write the entire essay in class with a pencil while I stare at you.

There are absolutely times when an in-class handwritten assessment makes sense. I use one as part of my own system.

But I don’t want my entire writing program designed around preventing misconduct. I want it designed around producing better writers.

So instead of making an assignment inconvenient, I’d rather make the student’s thinking visible.

Use the handwritten draft as a baseline. Require revision. Make those revisions easy to identify. Keep the drafting process in a document where you can review its history. Ask students to explain changes when necessary.

None of this makes AI misuse impossible.

But it makes it a whole lot harder to replace the student’s work with a generated essay without leaving obvious gaps in the process.

Let Students Talk About Their Writing

This might be one of the simplest tools we have.

Ask a student about their essay.

Not as an interrogation. Just talk to them.

“Why did you change this sentence?”

“What was wrong with your original thesis?”

“Why did you choose this quotation?”

“You used the word undermines here. What exactly is being undermined?”

A student who actually wrote and revised the piece can usually talk about the thinking behind it—even if the writing itself isn’t perfect.

And that distinction matters.

I would much rather have an awkward paragraph a student can explain than a flawless paragraph they don’t understand. One of those gives me something I can teach.

Students Still Need to Write

I think this gets lost in the AI conversation sometimes.

There’s a difference between acknowledging that AI exists and deciding students no longer need to develop the skills AI can imitate.

Students still need to learn how to make an argument. They still need to choose evidence. They still need to explain why that evidence matters. They still need to revise unclear thinking and make intentional choices about language.

AI can generate writing.

That is not the same thing as learning to write.

If anything, AI makes it more important that students understand the thinking underneath good writing. Otherwise, they have no way to judge whether a generated response is insightful, shallow, repetitive, inaccurate, or complete nonsense.

We Don’t Need to Panic. We Need Better Process.

AI is absolutely creating real problems in English classrooms. Students are going to misuse it. Teachers are going to get frustrated. Some assignments that worked beautifully five years ago simply will not work the same way now.

But I don’t think the answer is abandoning writing or retreating completely to pencil and paper.

I think the answer is designing assignments that leave a trail.

Draft. Type. Revise. Highlight. Review.

Let students use technology—but make their thinking visible along the way.

Because the goal was never really the five-paragraph essay sitting in the gradebook.

The goal was the student who learned how to think well enough to write it.