1

Rethinking the Writing Assignment in the Age of Generative AI

Liam Carter

First-Year Writing Program, State University

WRIT 101: College Composition

Professor Hadley

April 18, 2026

Rethinking the Writing Assignment in the Age of Generative AI

The college essay is dying, or at least changing in ways that look like dying to the

people most invested in its current form. Generative AI tools released since 2022 can produce

a competent five-paragraph essay in under a minute, and undergraduate students at virtually

every institution now have free access to these tools. The educational response has been

uneven and often contradictory: some instructors have moved aggressively toward AI-assisted

writing as a pedagogical opportunity, others have reverted to in-class blue-book exams, and

most are somewhere in between, hoping the question resolves itself.

This paper argues that the question will not resolve itself, and that the right educational

response is neither full embrace nor full prohibition. It is a redesign of the writing assignment

around the cognitive work that AI cannot easily replace: the development of an argument over

time, the integration of the writer's own reading history, and the documentation of the drafting

process itself.

The strongest argument for prohibition is that AI assistance prevents students from

developing the underlying skills the writing assignment is designed to teach. Writing, in this

view, is not a way of communicating ideas you already have. It is a way of finding out what

you think (Murray, 1972). Outsourcing the drafting process to a tool short-circuits the

cognitive work that produces understanding, and the resulting essay, however polished,

represents nothing the student actually learned.

This argument is stronger than its current advocates often realize, but it is not strong

enough to support a blanket prohibition. The same argument has been made about every

previous technology that reduced the cognitive friction of writing: the typewriter, the word

processor, spell-check, grammar-check, and Google Docs comment threading. In each case,

the prediction that the new tool would prevent learning turned out to be wrong, because the

learning shifted to other parts of the writing process. Spell-check did not eliminate the need to

think about words. It eliminated the need to remember which letters they contained, which is
 3

not the same thing.

The harder question is whether AI assistance is different in kind from these earlier

tools. The case for difference rests on the fact that AI can produce the entire essay rather than

assisting with subtasks. This is real, but it is also a feature of how the tool is currently being

used, not a feature of the tool itself. A pedagogy that assigns "write me a five-paragraph essay

on X" and accepts a finished product will be replaced. A pedagogy that asks students to

document their thinking process, defend their choices in conversation, and revise based on

instructor feedback will not be.

The right response to AI in the college writing classroom is neither prohibition nor

unconditional embrace. It is a redesign of the assignment around what the technology cannot

do: ask a student to defend their argument in real time, articulate why they made the choices

they made, and connect the current essay to the reading they did three weeks ago. These are the

components of writing that AI tools, at least currently, cannot replicate, and they happen to be

the components most worth teaching.

This redesign requires more work from instructors than either prohibition or embrace.

It requires actually reading the drafts, conducting individual conferences, and accepting that

the artifact at the end of the process is less important than the visible work of getting there.

That is, however, what the writing classroom was supposed to be doing all along. The arrival

of AI has clarified the difference between teaching writing and grading essays, and that

clarification is overdue.
 4

References

Murray, D. M. (1972). Teach writing as a process not product. The Leaflet, 71(3), 11-14.

Mollick, E. (2024). Co-intelligence: Living and working with AI. Portfolio.

Yancey, K. B. (2009). Reflection and electronic portfolios: Inventing the self and reinventing

the university. Conference on College Composition and Communication.