Photo illustration by Justin Morrison/Inside Higher Ed | History Skills and Wavebreakmedia/iStock/Getty Images
It’s 2030, and your English department graduate admissions committee is meeting. Perusing the applicant files, its members notice, maybe not for the first time, that each writing sample seems off. The main argument falters after a few pages. Ideas jump around in two- to three-page increments. Literary analysis goes just far enough to make a generally observable point before the essay moves on. Over all, the whole sample reads like a series of shorter pieces strung together under the umbrella of a general topic, but not a structured argument.
The committee doesn’t suspect the applicant used generative artificial intelligence to do their writing. There is a voice here, and, besides, why pursue a master’s or Ph.D. in English if you wanted large language models to do your work for you? But this paper—and the others populating all those files—just doesn’t sound like what graduate applicants used to write.
This future will be one of our own making as we professors erode the basics of humanities instruction and assessment to keep LLMs from turning writing assignments into opportunities to offload work. For the difference between this speculative situation and today results not from the use of LLMs in English and other humanities classes but from the alternative assignments professors are using to defend against artificial intelligence. What look like clever classroom innovations will inevitably undermine the skills that humanities education has treated as fundamental to its practice over the past century.
Prominent among these tactics is the turn (or return) to the handwritten assignment, whether it be in-class writing, essay scaffolding, journaling or blue book exams. Hannah Pittard writes in The Chronicle of Higher Education of the pleasure her creative writing students take in the slow work of writing by hand. The New York Times has proclaimed the resurrection of paper and pencil both in high schools and colleges as the workaround to AI cheating. The Wall Street Journal has reported on a surge in the blue book business, noting that humanities classes have abandoned the traditional typed essay.
In classrooms everywhere, we’re told, students are deciding if a word is worth the ink as they draft essays for 30 minutes at a time. Or they are cutting up their loose leaf and rearranging the sentences collage-style, as if coherence were just a matter of putting composition in the right order. These reports beam with self-congratulation: Our students still yearn to express themselves, but they are so busy, and the AI temptation is so great, that we have reinvented assignment and assessment to give them back themselves.
Stephen Marche cheekily remarks in The Guardian that when 86 percent of college students admit to using AI regularly, you know that 14 percent are lying. And if this means that the craft of writing is “dissolving,” as he puts it, then making students put pencil or pen to the page becomes the key to preserving the form of communication that, Plato notwithstanding, has defined education for millennia.
In-class writing and handwritten journaling have not been the only tactics for sidestepping AI. Creative assignments that seem too personal for LLMs to mimic have been rising. Commonplace book projects, in which students reflect upon notable passages, or direct responses that require students to put their own lives in dialogue with the readings, would be too meaningful, one hopes, for students to want to outsource. With an eye to older academic traditions, face-to-face oral examinations, whether with one student or with small groups at a time, offer the chance to evaluate students without the presence of screens.
None of these is foolproof, though. I have certainly received LLM-generated creative and personal assignments, and students can prepare for exams by having AI summarize texts and outline important points regardless of how deeply they read the texts and pay attention in class. My point is not to cynically discourage these new forms of evaluation so much as to point out that we are changing our assignments to solve a problem that, without student buy-in, we are not solving. And recent reports citing examples from universities as academically selective as Columbia and Stanford Universities, the University of Notre Dame, and New York University certainly suggest that student buy-in is, to put it mildly, not a sure thing.
Rory Truex writes that his Princeton University students appreciate the return to cramped hands and in-person work, but the assignments he describes—the change from a take-home, open-book exam in which students had eight hours to answer several essay questions, to an in-class, closed-book, three-hour handwritten final—reflect my point. Students may like proctored, forcibly timed and thus shortened work, but they don’t even know what they lose when long-form thinking and writing are banished.
A new classroom status quo stitched together to account for LLMs (subscriptions to which are paid for, in many instances, by the universities themselves) will have consequences beyond any individual classroom environment. These down-the-road effects remain undiscussed as everyone acclimates to the new normal of preparing students to be “AI fluent,” a concept that continues to be (let us generously say) in flux. And as humanists try to preserve a way of learning that has worked for decades without such fluency, their good-faith efforts will lead inevitably to the scenario I began with: Incoherent essays defined by cut-off analysis is the endgame of undergraduate humanities classes that chiefly revolve around short-form, handwritten assignments completed 50 or 75 minutes at a time, some fraction of two or three days a week.
In-class writing remains the “five-minute abs” of essay construction. We are condensing serious consideration over time of a book, poem or concept into five to 10 minutes of pondering a professor-generated prompt to be followed by 30 to 40 minutes of writing. Revision be damned, unless we sacrifice an additional class period so that 50 minutes of writing might be redone in … another 50 minutes. In a 15-week semester on a Monday/Wednesday/Friday schedule, how many teaching and discussion days are devoured by these writing days? To reach the bygone standard of 20 typed pages a semester it could, conceivably, be most of them—an absurd conclusion.
It is a truth universally acknowledged that the undergraduate essay has little relevance to the world beyond college. Nowhere are the jobs that require a paper, anywhere from five to 15 pages, written in isolation for an audience of one to prove that you have digested something you’ve read for no immediately practical reason. Nonetheless, the undergraduate paper as it has been practiced for decades did signal something, both in general and to future employers: the student’s willingness to engage in the slow consideration of another’s writing and ideas and then respond to them over many hours of targeted attention. Writing a successful paper means rejecting one’s cursory reactions and forcing oneself to take and defend a position that doesn’t even appear for the (at minimum) first 50 minutes of consideration.
Papers are the lab work of the humanities. They are the crucible for ideas invented, discarded, reanimated, rearranged and revised into sentences that reveal something to the writer that they didn’t know before. Papers communicate in ways that a 10-minute conversation or a 30-minute wrestling match with loose leaf can’t, because writing and rereading and rewriting and rethinking is fundamentally not a 10-minute or 30-minute activity. Has every student always done this? No. But I challenge anyone to find me the undergraduate who has produced a worthwhile (or even just honestly passable) six- to 10-page paper in 50 minutes, or even 100.
No longer will an English or history or philosophy degree deliver on the promise of dedicated thinking and working sustained over time. No longer will these or similar degrees indicate that a student has, for a third or more of their credit hours, practiced and perhaps even mastered the work of assembling sentences into paragraphs that build on each other to forge a coherent argument. Is every paper a success, or is every paper worth reading by anyone other than the professor? Of course not. Yet every paper is, or has been, written—if not painstakingly, then at least without the promise of instant gratification. We delude ourselves thinking that “papers” produced in 50 minutes and evaluated on little more than a check/check plus/check minus system (regardless of the letter grades we use as masks) even aspire to this erstwhile standard.
In preserving the humanities from AI shortcuts and feeling good about what happens in the classroom, we are destroying what the degrees mean. Jonathan Kramnick has put it concisely by proposing that close reading is “an expert practice of writing,” a craftwork that requires weaving one’s words together with those of others. It cannot happen orally; it cannot happen in fits and starts; it cannot be begun and concluded in 50 or 75 minutes. The undergraduate essay, if it signified anything, signified that students had immersed themselves in this difficult exercise and could carry that effort into the world. We cannot promise this experience anymore.
I began with a scenario concerning graduate admissions, a procedure that affects a small minority of college students and is seemingly less alarming than the hollowing out of the hard work of humanities education. Yet we threaten the very continuance of our own profession when we fail to prepare potential graduate students for what to expect. Academic writing is not a handwritten, blue book–encased exercise. Nor is it done incrementally under the watchful prompting of a teacher or proctor. Will advanced classes include longer, at-home papers so that ambitious students have occasions to practice more extended writing? Or will these advanced classes be bogged down in basic instruction in classic essay writing, a genre soon to be nonexistent in lower classes, not to mention high school? Or will professors choose which students get the chance to write longer essays so that they might practice before moving on to graduate education? Or will everyone get the option to write in-class or at home? Can we know which students will genuinely do the work, and which ones merely want to do the work but are open to shortcuts, and which ones will choose the at-home assignment and use LLMs with no misgivings?
I can’t answer any of these questions with certainty (nor can anyone else, based on recent reporting), and the LLMs our students use in the eternally recurring present that is university education will always be the worst version of what is to come. In other words, deciding how to assign and evaluate papers for students who aim for graduate school is never going to get easier. Meanwhile, our solutions to the LLM crisis in the humanities ensure that fewer than ever will want to move on or even be capable of doing so, at least according to the standards we have hitherto upheld. For those few who do successfully become professors, their experiences of short-form coursework upon which they construct their pedagogical and research practices are unlikely to ever lead undergraduate education back to the essay. Eventually, even the option will disappear.
Whether our undergraduate humanities majors plan to continue to graduate school or not, what we are able to ask of them no longer aligns with what we have assumed a humanities education entails. There may not be a solution in a world of cheap, accessible LLMs that every incoming undergraduate will have been using since middle school, if not before. In that case, what it means to work in the humanities will have to change, and if that is the case, we need to take a hard look at the legacy of 20th-century classroom and research practices. If we truly believe that handwritten, in-class papers or oral examinations or creative assignments record proficiency in a course and competency in long-form thinking and writing, why have they been minimized for decades? Have we been incorrect in promoting the undergraduate essay as a worthwhile project?
I do not believe so, but the about-face witnessed across universities since the penetration of AI has not been paired with rigorous debate over what a humanities education will signify once we behead the practices of the past. It cannot remain the same, but are professors prepared for what slouches toward us to be born? Or if we do decide it can remain the same despite a complete overhaul of the foundations of humanities work, then what are we saying about our long commitment to the value of essay writing as a time-intensive, cognitively sophisticated experience?
The alternative assignments praised as workarounds to the laziness of LLM dependency will prove to be not solutions but stopgaps. Will our degrees just come to signify that students have mastered exercises constructed primarily to circumvent artificial intelligence? Humanities degrees may never have funneled students seamlessly into the job market, but they certainly meant that a level of higher-order thinking had been practiced.
Perhaps no longer. I fear that our cures for the weakening of humanities education in the face of probability engines will turn out to be the poison that finally ends the humanities profession as we have practiced and defended it.
