I remember the moment I knew my approach to student use of artificial intelligence was not working.
Early in a meeting at N.Y.U.’s Abu Dhabi campus last fall, a philosophy professor, arms crossed over his chest, told me he’d tried one of the strategies my office had suggested — talking with his students about the ways A.I. could interfere with their learning — and it hadn’t worked. His students had listened politely, then several of them had used A.I. to write their papers anyway. He particularly wanted me to know that “even the good students,” the ones who showed up to class wanting to talk about the readings, were using A.I. to avoid work outside class.
This was a theme I’d hear over and over, listening to faculty members across disciplines at the end of the semester; even some of the students who obviously cared about the material and seemed to like the classes were no longer doing the hard work of figuring out what they wanted to say. Our A.I. strategy had assumed that encouraging engaged uses of A.I. — telling students they could use software like ChatGPT to generate practice tests to quiz themselves, explore new ideas or solicit feedback — would persuade students to forgo the lazy uses. It did not.
We cannot simply redesign our assignments to prevent lazy A.I. use. (We’ve tried.) If you ask students to use A.I. but critique what it spits out, they can generate the critique with A.I. If you give them A.I. tutors trained only to guide them, they can still use tools that just supply the answers. And detectors are too prone to false accusations of cheating and too poor at catching lightly edited output for professors to rely on them.
Learning is a change in long-term memory; that’s the biological correlate of what we do in the classroom. Now that most mental effort tied to writing is optional, we need new ways to require the work necessary for learning. That means moving away from take-home assignments and essays and toward in-class blue book essays, oral examinations, required office hours and other assessments that call on students to demonstrate knowledge in real time. The shift is already happening: The Wall Street Journal reported on booming sales of blue books last school year.
Students and teachers alike are skeptical of these changes. One professor I know described her new reliance on in-class work as “teaching high school.” But these strategies do not represent a loss of rigor. They are simply a return to an older, more relational model of higher education.
Talking, listening and reading have been part of academic culture since the beginning, but written assignments — the five-paragraph essay, the research paper, reading responses — were not. In the earliest universities, which coalesced in a handful of European cities around a thousand years ago, books were scarce, movable type was nonexistent, and education was organized around oral instruction and examination.
The words “lecture, lecturer” and “lectern” all derive from the Latin verb “lego,” which typically means “to read.” In the medieval university, lectures involved teachers reading to the students from a book, sometimes the only copy the institution had. Sometimes professors added their commentary to the reading; sometimes they did not. Some students were expected to write down what they heard, some merely to listen. At times, writing was discouraged. In 1355 the arts faculty at the University of Paris forbade masters to lecture at a slow speed that would have allowed students to copy their words verbatim.
You can still see traces of that old academic culture in Ph.D. programs, in which students have to pass oral exams and defend their thesis in a viva voce (“with the living voice”) in conversation with their examiners. Cambridge and Oxford, the inspiration for most early U.S. colleges, did not meaningfully adopt written exams until the 18th and 19th centuries, half a millennium after they were founded. The shift to original, written student work was partly in response to instruction in increasingly technical fields and partly due to the fact that written work made it easier to teach more students.
Even in the U.S. our earliest colleges followed the tradition of oral examinations. Emphasis on students writing compositions did not spread until we started copying German research universities in the 1870s. Freshman comp, the standard U.S. writing class, shifted to expect more unique and expressive content from students after World War II.
All of which is to say that our current practices around student writing are not part of some ancient tradition. Which assignments are written and which are oral has shifted over the years. It is shifting again, this time away from original student writing done outside class and toward something more interactive between student and professor or at least student and teaching assistant.
Though the return of the blue book exam is one sign of this change, a number of older practices for assessing student learning are being revived. Faculty members can engage students in conversation, via Socratic dialogue or simple Q&A. They can cold-call students or get students to ask one another questions. This may require device-free classrooms, as many faculty members also report students using ChatGPT in class to generate answers to questions posed in class.
Faculty members can require students to come to office hours, have similar unscripted interactions or ask them to perform tasks that they can manage only if they’ve learned the material (called authentic assessment). We can ask students to write something in class one week and revise it in class the next. There are also tools for remote proctoring or locking down browsers to prevent students from using A.I. while they work even if they are not in a classroom. Some schools are building internet-free classrooms for holding exams. The goal is to get the students to demonstrate what they have internalized from their work.
There is the problem of scale. With some lecture classes in the hundreds of students, in-class conversation is a nonstarter. We may have to do more assessment in sections run by teaching assistants or hold proctored, in-person writing sessions. Studies testing large-scale oral exams are already being published.
I have talked to hundreds of N.Y.U. faculty members over the years, and few of them like this. Some of this feeling is just the annoyance any people would feel having their work routines disrupted — rethinking assignments and reworking syllabuses, along with considerable restructuring of class time. More important, though, this shift to extemporaneous and oral performance means losing the ability to give students moderately complex goals that they have to wrestle with on their own. Timed assessment may benefit students who are good at thinking quickly, not students who are good at thinking deeply. What we might call the medieval options are reactions to the sudden appearance of A.I., an attempt to insist on students doing work, not just pantomiming it.
Students tell me they don’t like the new forms of assessment, either. Some of this is just the annoyance anyone would feel at having a labor-saving tool taken away. But some students may struggle with the practicalities of this new system. This is a generation that never learned cursive; its members grew up typing. For many of them, timed essays are not a return to anything but a new and unfamiliar mode. Some are already so reliant on A.I. that working without it is disorienting, even upsetting. The student cohort entering college this fall will have had access to generative A.I. for most of their high school years. A colleague at another university recently reported a student saying, of a return to in-class exams and limited device access, “It’s like they want us to fail.”
We, of course, do not want our students to fail, but we don’t want them to fail to learn, either. A student who cuts and pastes a history paper is enrolled in a cutting and pasting class, not a history class. If the student’s preferred working methods reduce mental effort, we have to reintroduce that effort somehow.
Tuesday is the 1,000th day since ChatGPT’s release. In that short time, it’s already clear that the arrival of software that can generate unlimited amounts of OK-ish text will devalue many kinds of writing. There will still be a market for quality, just as there is still a market for films, even though TikTok exists, but the production of ordinary writing now requires much less skill. As business writing gets automated, university writing will return to its historical state, with more emphasis on students writing to commit things to memory, rather than to create a written artifact.
Over time, though, we will adapt. (Constant incremental adaptation is how colleges and universities work.) Despite frequent pronouncements that college is doomed because students can now get an education from free online courses or TV or radio or the printing press, those revolutions never flattened us. Nor will A.I. Contrary to much popular opinion, college is not in the information transfer business; we are in the identity formation business. Our medieval turn will not be a wholesale reversion. Blue books and viva voce testing will live side by side with modern innovations like active learning and authentic assessment. But a return to a more conversational, extemporaneous style will make higher education more interpersonal, more improvised and more idiosyncratic, restoring a sense of community to our institutions.
Clay Shirky, a vice provost at N.Y.U., has been helping faculty members and students adapt to digital tools since 2015.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.
The post The University’s Best Weapon Against A.I.: The 14th Century appeared first on New York Times.
