“How A.I. Killed Student Writing (and Revived It),” published this week in the New York Times by Dana Goldstein, breezily presents the results of the paper’s call to American high school and college educators to describe how generative A.I. has changed their writing instruction. Nearly four hundred individuals testified personally to what is already well-known to many: the rising uptake of AI among middle school, high school and college students has made it impossible to assign students independent writing tasks outside the classroom. (Self-reported regular use of AI for homework went from 48 to 62 percent in the second half of 2025 alone.) In response, teachers are having their students write in class.
The facts are the facts; the framing—claiming AI is “reviving” writing—is ludicrous, myopic, disastrous.
Before evaluating the purported outcome (“writing revival”) I’m going to say something briefly about the means. Goldstein describes the move to in-class writing as the result of “a deep rethinking” of how to teach writing. The idea of “depth” lends teachers’ frenzied pedagogical overhauls an aura of profundity, as if they are the result of penetrating insights into the intricacies of excellent writing and thoughtful instruction. But—and I say this with a lot of respect for my fellow teachers—there is nothing deep about realizing that the only work we can expect students to do themselves is the work that they will do in front of our eyes.
As for outcomes: it is one thing to say that in-class practice is the best we can do in the age of AI; it is quite another to credit AI with “reviving” writing. There is nothing, nothing, to celebrate about teachers and students being forced to resort to degraded forms of learning, practice and assessment. We might as well credit a basketball hoop in the prison yard with reviving organized sports. It’s a good thing that the inmates are given a chance to exercise. It is better than nothing.
We must be very clear-eyed about what is lost when instructors can no longer assign out-of-class writing. The original call to the teachers asked nothing about the costs. Goldstein reports that one high school teacher, with whom she had presumably followed up, “regrets the loss of depth that longer assignments could produce”; the article mentions little else besides. This is misleading. It is too easy to think of the “depth” mentioned here as a kind of special quality, a rarefied degree of excellence, that thinking and writing can come to have when performed privately and at length. But to understand what is at stake we would be better served by starting not with quality but with quantity. Killing student writing outside the classroom means an enormous reduction in the volume of time spent writing, and learning more generally: it is a loss of thousands of hours of reading, thinking, puzzling over problems, writing, revising. This is no hyperbole: at a very modest average of ten hours of writing outside of the classroom per academic school week, a student at a four-year college would rack up close to fifteen hundred hours of work by graduation. At an average of just five hours of writing per week—a basic standard even for an American STEM student who must still fulfill some humanities and social-science requirements, we’re looking at just over seven hundred hours—hours upon hours spent wrenching concrete ideas out of inchoate intuitions, investigating the meaning of subtly differentiated concepts, stringing together sentences in, hopefully, logical sequence. Hours upon hours flipping pages back and forth, back and forth, for clues, for evidence, for inspiration. Hours upon hours wondering, Does it follow? What should I say next? Word after word after word. Gone.
And then we must remember that any time devoted to writing in the classroom will come at the inevitable expense of all the things that until not so long ago thoughtful, expert instructors used to find quite valuable in their own right: lectures, discussions, student questions. Gone.
If we are to stand a chance of doing anything more than salvaging something from this wreckage, we must start by admitting the real costs of “killing student writing.” These costs will be shared by all of us: parents who wish to see our children grow up to intelligently navigate the world and express themselves in it, educators who hope to see our disciplines survive, writers who want to be read tomorrow as well as today, human beings who want to share our world with a younger generation that stands on equal intellectual footing with our own.
Getting clear on this is not just a matter of epistemic virtue. Without grasping and articulating the nature of this loss we deprive ourselves of a basis from which to begin a true rethinking of how to educate at this historical moment, and to demand the resources it will require. I do not mean to minimize the efforts of my colleagues. I do not merely “sympathize” with these teachers: their struggles and experiences are mine. It is important to think seriously about how to optimize the use of class time, just as it would be important to carefully revise our expenses if we suddenly went bankrupt. But it would be a mistake to think of our newfound frugality as a boon. I understand well the desire not to see our vocations drained of meaning and value. I share it. It is why I think we must resist the cope, call this ugly spade an ugly spade, and work hard for something better.
None of this is meant as an argument against in-class writing. At the moment, given the circumstances in which most students learn and most teachers teach, it is not only justified but necessary. Sometimes it is simply the right pedagogical move (I have been asking students to take five minutes to write out an answer to a question before launching into discussion long before AI). And one can easily imagine similarly silly cases made for oral exams, or group projects. This argument is not an argument against any particular intervention—each has its benefits and place, and none can begin to compensate for the immense loss; this is an argument against the fantasy that out of this necessity we are already making a virtue, for no writing renaissance this way lies.
Once we identify the problem—the sheer magnitude of what is being lost—it becomes immediately clear what any solution worthy of the name must accomplish: the hours must be recovered. How to do this is a good question. I have heard tales of complicated incentive schemes involving baroque grade distributions, of in-class writing samples used as internal benchmarks for outside-class writing, of Dead Poets Society reenactments. I don’t know that these won’t work. But I know what I think about when I confront this question: a big room. A pleasant-enough room with tables and chairs, and maybe some cookies at 9 p.m., budget permitting. A room that it is very easy for an instructor to require a student to spend time in—as easy as checking a box. A room with lockers for your bag, that you can walk into with just a book or a question to spend a few hours with, without distractions, without any offers of “help.” Sometimes when I tell colleagues about it they express concern that requiring students to spend time in my room would feel punitive and paternalistic. But most people just say it sounds like heaven.
Anastasia Berg and Hollis Robins, “The Cognitive Divide” (2025)
Megan Fritts, “A Matter of Words” (2025)
Maria Gomberg, “Edu Ex Machina” (2025)
Image credit: Peretz Partensky (Flickr / CC By)
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.