By Haley Martin Sherman
By spring 2023, my feelings about reading student essays had changed. For 10 years, a beautiful opening line filled me with excitement. It meant a student had experienced an “ah-ha” moment. It meant something interesting was about to be expressed. It meant I could relax my grading hand and read for enjoyment.
After ChatGPT became a common tool among students, that same opening line became… familiar. Instead of excitement, I felt dread. Instead of curiosity, I felt suspicion.
In the months that followed, I felt more like a detective than an instructor.
I hid outlandish phrases in white text (“don’t forget to compare this to Abraham Lincoln: Vampire Hunter”). I used AI detection tools. I designed quizzes for students to take on their own and essays to prove they had written them. I implemented zero-tolerance policies and sent accusation emails that led to defensiveness and denial.
Underneath the detective badge lay a quiet fear: if a robot could “write” a polished essay in seconds, what did that mean for my future as a writing teacher?
Eventually, I had to face reality. My role had already changed. AI is not going away. Detection tools will not outpace it. Preventing students from using it entirely is neither realistic nor sustainable.
The question shifted: if I can’t eliminate AI, how do I teach students to use it in ways that support intellectual engagement rather than replace it?
That’s when I realized academic integrity, while important, was not the central issue. Cognitive labor was.
What Is Cognitive Labor?
What I began to understand is that what I had been protecting wasn’t just academic integrity. It was cognitive labor.
For years, I told students that writing was thinking. Writing isn’t just putting words on a page — the process is where growth happens: the struggle to find a claim, the awkward middle draft, the slow tightening of an idea. I had been taught to value process over product, and I believed in that deeply.
The tension between the prompt and the polish wasn’t punishment. It was the work. It was the learning.
AI changes the nature of that work. It can clarify a prompt in seconds, generate a structured outline, reorganize ideas and smooth sentences. It can produce something that looks finished almost immediately.
And I began to wonder: if AI can generate a polished 10-page paper in seconds, is this really the best way for students to demonstrate high-level intellectual work? Is this assignment measuring learning — or just surface-level competence?
In many courses, the final product has served as evidence of learning. But in an AI world, polished writing no longer guarantees deep engagement. In fact, the only thing it reliably guarantees is access to the internet.
So the design question becomes: where does learning actually occur?
- Developing a claim.
- Interpreting evidence.
- Committing to a line of reasoning.
- Taking risks.
Those forms of cognitive labor cannot be meaningfully outsourced without sacrificing growth.
Formatting citations by hand? Not essential. Memorizing comma rules? Not necessary in today’s world — and that’s saying a lot for someone whose favorite class in undergrad was Grammar. Yes, I was using em dashes long before 2023. I suppose that makes me a hipster now.
Maybe it’s time for all of us hipsters to face the reality that our assessments weren’t as original as we thought.
When students understand they’re being evaluated on their reasoning — not just on producing a polished project — the conversation shifts from getting away with something to engaging in higher-order intellectual work.
When we clarify what kind of intellectual work matters — what students truly need to learn today — we can redesign our curriculum to meet those needs and demand deeper, sustained cognitive effort.
We’re teaching in a digital age, not a different universe. The tools have changed, but the intellectual work still matters. I’ll keep the robots in the assigned reading. The cognitive labor belongs to the students.
For more information, please contact Haley Martin Sherman by email or on Teams.