 |
| They Refused. Here's Why. |
Every week brings a new story about AI transforming the classroom—universities signing deals with tech companies, courses redesigned around chatbots, workshops promising to make every professor "AI-literate." The message is consistent: adoption is inevitable, and the only question is how fast you adapt.
But a growing number of academics have quietly reached the opposite conclusion. They have read the pitch, tested the tools, weighed the costs—and decided to say no. Not out of fear or nostalgia, but out of conviction. This is their story, and their numbers are larger than most headlines admit.
Not a Fringe Position Anymore
In July 2025, more than 1,400 educators from around the world signed an open letter refusing "the call to adopt GenAI in education." Their language was blunt: they called generative AI "a threat to student learning and wellbeing," driven by "a massive marketing push" with "insufficient evidence" of real learning gains. Signatories pledged not to use AI for marking, course design, or rewriting curricula around "AI literacy."
The resistance is not only individual. In March 2026, the Conference on College Composition and Communication—the largest professional body of writing educators in the United States—passed a resolution affirming the right of students and teachers to refuse generative AI in the writing classroom, framing the choice as a matter of academic freedom rather than a disciplinary matter. The resolution also took aim at the assumptions beneath the marketing: unsubstantiated productivity claims, data privacy erosion, and the quiet conversion of student work into training data for corporate models.
Refusal, in other words, is no longer one eccentric professor muttering in the faculty lounge. It is an organized position, with documents, signatures, and a professional community behind it.
The Reasons Run Deeper Than "Cheating"
Ask an AI-refusing academic why they refuse, and plagiarism is usually the fifth reason they mention. The first four are structural.
- Learning is the work. If an AI writes the draft, the student never performs the struggle—the drafting, the false starts, the slow formation of judgment—that writing exists to produce.
- Labor and consent. These models were trained on academics' own writing, often without consent or payment. Many professors see teaching the tool as subsidizing the machine that is being positioned to replace them.
- Reliability. The tools invent sources, fabricate citations, and smooth errors into confidence. For disciplines built on evidence, that is a disqualifying flaw, not a bug to be patched.
- The environment. Every query has a cost in energy and water, and several signatories cite the climate footprint as part of their refusal.
Strip away the jargon and the position is simple: a technology that is unreliable, built on unconsented labor, harmful to the planet, and corrosive to the skill being taught is not something a responsible educator adopts by default.
What the Surveys Actually Show
The adoption narrative says everyone is using AI. The data says otherwise.
An Ithaka S+R survey of postsecondary instructors found that 42% completely prohibit their students from using generative AI in their courses—including 53% of humanities faculty and 41% of social scientists who have not engaged with AI in their teaching at all. At the University of Maryland, 86% of instructors described themselves as non-routine users of generative AI for instructional work. A faculty survey at Metropolitan State University of Denver found 61.5% of respondents had never used generative AI tools for educational purposes at all.
Read those numbers again: in multiple independent surveys, somewhere between half and nine-tenths of faculty are not routine AI users. The classroom has not been transformed. It has been pitched to—and a large share of its residents are declining the offer.
The Classroom Is the Counterargument
Refusing academics rarely just delete the topic. They rebuild the classroom around the reasons for their refusal.
Writing professors have revived blue-book exams and in-class drafting sessions. Seminar leaders have moved back to oral exams, where a student must defend an argument in real time. Historians assign work with primary documents that resist summarization by a chatbot. Computer scientists who once taught "coding" now teach debugging by hand—precisely the skill an autocomplete erases.
They are not teaching around the AI. They are teaching through the gap it leaves.
Their shared logic: the moment a task can be fully outsourced to a machine, the task itself is the wrong assignment. The goal was never the finished essay; it was the mind that changed while writing it.
"Inevitability" Is a Marketing Line
The strongest pressure on refusing academics is the claim that resistance is futile—that AI is coming anyway, so holding out only disadvantages students. The open letter's signatories have an answer: the inevitability narrative is itself the product.
It is a convenient argument for companies whose business model requires universities to buy in, and for administrators whose adoption decisions look better when described as destiny. History is full of "inevitable" technologies that education absorbed slowly, selectively, or not at all. Teaching did not adopt every innovation industry offered it; it negotiated with them, and sometimes it won the negotiation.
Refusal, in this reading, is not standing in front of the train. It is asking whether the train is going anywhere students need to be.
Students Are Refusing, Too
The movement is not only about professors protecting students from AI. In many classrooms, the students got there first.
Some students refuse generative AI as a matter of pride: why outsource the exact skill they enrolled to learn? Others resent the hypocrisy of institutions that warn against AI in submissions while adopting it for marking and administration. The CCCC resolution explicitly extends the right of refusal to students, rejecting the assumption that a student who declines AI is lazy or afraid of technology.
A refusal movement that ignores students is paternalism. This one, notably, has students inside it.
The Cost of Saying No
None of this is free. Refusing academics pay a real price.
They are described as "behind the times" in faculty meetings. They lose out on AI-related grants and teaching awards. They watch course-release incentives flow to colleagues who adopt, while their own refusal earns them extra preparation work—because human-centered teaching is slower and more labor-intensive. Junior faculty, in particular, feel the squeeze: saying no is riskier before tenure.
There is also the question of students. Some undergraduates arrive expecting AI to be part of the course, and a refusal policy can read as an inconvenience rather than a philosophy. Refusing academics spend a lot of time explaining the why—in syllabi, in class, in office hours—because a refusal without a reason looks like a rule, and rules without reasons get broken.
Honesty requires saying that refusal is a privileged position in some respects. Adjuncts with five courses and no job security have less room to refuse institutional mandates than a chaired professor does. The open letter's commitments are easier to keep with a stable contract.
And yet the signatures keep coming—which suggests that for many, the cost of adopting has quietly become higher than the cost of refusing.
The Middle Ground Is Real
It would be easy to paint this as a war between purists and enthusiasts. The reality is messier, and more interesting.
Many of the refusing academics use technology constantly—databases, statistical software, digital archives. What they refuse is specifically generative AI: the tool that produces the thinking for the student, the draft for the author, the feedback for the teacher. Some happily use AI for their own email and administrative chores while banning it from anything connected to learning. Others allow students to use it for brainstorming but not for drafting.
The common thread is not technology-hatred. It is jurisdiction: a refusal to let the machine into the part of education where the human is the point.
Where Each Camp Usually Lands
The full refusers: no generative AI in any part of their teaching—not for marking, course design, or student work.
The classroom refusers: personal use is fine; the classroom stays human—no AI-graded work, no AI-generated feedback, no AI-written student submissions.
The conditional adopters: AI is allowed for narrow tasks—brainstorming, translation help, editing—with disclosure required and the final work demonstrably the student's own.
The enthusiasts: AI is integrated throughout, often with the argument that students must learn the tool they will use at work.
All four camps exist on most campuses. What has changed recently is that the first two have stopped apologizing.
The Bottom Line
The academics refusing generative AI are not relics defending a photocopier against the internet. They are, in many cases, the people who have thought longest about what teaching actually is—and who have concluded that a tool which does the thinking for the learner is not a neutral addition to the classroom but a subtraction from it.
Whether they are right will be settled slowly, classroom by classroom, in the quality of the minds that emerge a decade from now. But one claim can be settled today: the refusal is real, it is organized, it is growing, and it deserves to be taken seriously—not as a tantrum against progress, but as a serious argument about what progress in education should mean.
This article is for general information only and reflects positions drawn from public surveys, resolutions, and open letters cited above. It does not constitute academic, legal, or institutional advice.
No comments:
Post a Comment