If machines can write essays, what’s the point of assigning them?
That question, raised by Hua Hsu in his New Yorker essay What Happens After A.I. Destroys College Writing?, lingers like a quiet challenge to every educator who still grades student papers late into the night.
Hsu presents a picture of college life where ChatGPT, Claude, and Gemini are as common as highlighters once were. Students use them to summarize readings, organize notes, write essays, and even craft text messages. To many of them, this isn’t cheating, it’s just the next logical step in academic survival. AI, in their view, has become the new calculator for the humanities.
Hsu’s story follows students who speak with surprising candor about how they use AI to breeze through courses that don’t inspire them. They aren’t trying to deceive anyone; they’re trying to manage an overloaded system that values grades and deadlines more than curiosity and struggle. Professors, meanwhile, scramble to restore authenticity, reviving in-class essays, blue books, and handwritten exams, hoping that a pen can somehow stand against the algorithm.
But beneath the anecdotes, Hsu’s piece cuts deeper. He hints that maybe the real issue isn’t AI at all. Maybe the problem is how we’ve defined learning for so long. We’ve built assessment systems that reward the finished product, the essay, the report, the polished paragraph, while ignoring the process that leads there. When an AI tool produces that polished product in seconds, it simply exposes how much of our grading has been tied to appearances, not thinking.
That’s the uncomfortable truth AI has made visible. The tools didn’t invent shortcuts; they revealed that our system has been one all along. For decades, the academic essay has stood as the ultimate proof of learning, a single document meant to represent weeks of reading, reflection, and revision. Yet most teachers know that the real thinking happens long before the final version: in the messy drafts, in the moments of confusion, in the quiet effort to connect one idea to another.
AI didn’t destroy the essay. It exposed how fragile our faith in it was.
If the goal of education is to nurture independent thinkers, then we need to stop grading the end point and start valuing the journey. Learning has always been iterative. It happens through drafting, feedback, rethinking, and rewriting. Students grow when they wrestle with uncertainty, not when they outsource it. The challenge ahead isn’t to ban AI from classrooms but to rebuild assessment so that the human parts of learning (i.e., voice, reasoning, reflection) stay visible.
That means shifting our focus from what students produce to how they produce it. A short essay written entirely by hand, followed by a brief reflection explaining how it evolved, can tell a teacher more about a student’s mind than a polished ten-page report typed overnight.
A class discussion that requires students to defend an idea aloud reveals depth of understanding that no AI can fake. When teachers ask students to connect lessons to their personal experience or to local issues, they make thinking contextual and alive again.
In that sense, AI hasn’t ended writing, it has given us a reason to rethink why we assign it in the first place. What if essays weren’t just proofs of knowledge but spaces for intellectual risk? What if grading rewarded persistence, curiosity, and revision instead of fluency and polish? These are the kinds of questions Hsu’s piece invites us to ask.
AI will keep improving. It will get faster, smarter, and better at sounding human. But education doesn’t have to compete with that. Its strength has always been in the human struggle to make meaning, to understand, question, and create. Machines can produce text; they can’t experience the slow discovery of thought.
If we want assessments that survive the age of AI, we have to build them around that discovery.
Related: Top AI Lesson Plan Tools for Teachers
6 Strategies for AI-resistant Assessments
The following strategies offer practical ways to create AI-resistant assessments and keep authentic learning at the center of classroom work.
1. Make learning process-oriented
Shift focus from polished products to the thinking behind them. When students know their early drafts, outlines, and reflections count, they start paying attention to how ideas take shape. You begin to see their reasoning evolve, not just their editing skills. This approach also gives quieter or struggling students a chance to show growth, not just a snapshot of performance at the end.
2. Bring back in-person conversations
Pair writing with conversation. A short class discussion after an assignment can reveal how students understand their own work and how they respond when ideas are questioned. You can hear what’s authentic and what’s rehearsed. These exchanges help build confidence and accountability—the kind that comes from speaking your thoughts out loud and realizing your words carry weight.
3. Redesign assignments
AI thrives on vague or formulaic prompts. When tasks are specific—rooted in classroom readings, personal connections, or community issues—the generic AI voice starts to crumble. Ask students to apply a theory to a local event, or connect a concept to their lived experience. The goal is to make the task feel meaningful enough that only the student in front of you could have written it.
4. Diversify how students show understanding
Writing is one way to think, but not the only one. Invite students to express what they’ve learned through podcasts, short videos, collaborative projects, or visuals. These forms test comprehension and creativity in ways that automated text tools can’t reproduce. When students choose how to present their ideas, they also take more ownership of them.
5. Keep Writing Human
AI can draft and polish, but it can’t think the way people do. Writing has always been about reasoning, arguing, and shaping a personal voice. When students wrestle with words, they’re learning how to clarify ideas and test perspectives. That’s the real purpose of writing—the slow, sometimes messy process of thinking on paper. AI only makes that purpose more visible, reminding us why human expression still counts.
6. Be Clear About AI Boundaries
Students need to know where AI fits in and where it doesn’t. Share a simple “traffic light” scale: green for tasks where AI can help (like brainstorming or grammar checks), yellow for partial use with teacher approval, and red for assignments that must remain fully human, such as reflections or argument-building. When these boundaries are clear, students learn to treat AI as a guided partner, not a shortcut.
References
- Hsu, H. (2025, June 30). What happens after A.I. destroys college writing?The New Yorker.https://www.newyorker.com/magazine/2025/07/07/the-end-of-the-english-paper
- Perkins, M., Furze, L., Roe, J., & MacVaugh, J. (2024). The AI assessment scale (AIAS): A framework for ethical integration of generative AI in educational assessment. Journal of University Teaching & Learning Practice.https://doi.org/10.53761/q3azde36
