In November 2025, the introduction to a forum in The Chronicle of Higher Education called the release of ChatGPT in 2022 “a clear inflection point in the history of higher education.” As a historian, I’ve come to think of it more as a nuisance, meaning a technological development beyond my control that I have to consider when designing the policies and practices that make up each syllabus and all of my assignments. Certainly, ChatGPT is not “a clear inflection point in the history of higher education,” because with just a few adjustments professors can continue to teach with largely the same methods and goals that they have always had. Nevertheless, too many university administrators and faculty members, deceived by rhetoric that originated in Silicon Valley, have overreacted to the arrival of artificial intelligence.
I don’t blame them or Silicon Valley entirely. The recent history of American higher education has in a sense been one hype cycle after another. In 2006, the sociologist Adam Best wrote, “Over the years, I have been assured that our university—if not all of higher education—was about to be transformed by the Pacific Rim, assessment, active learning, cooperative learning, distance learning, service learning, problem-based learning, responsibility-based management, zero-based budgeting, broadening the general-education requirements, narrowing the general-education requirements, capstone courses, writing across the curriculum, affirmative action, multicultural education, computer networking, the Internet, water (don’t ask), critical thinking, quantitative reasoning, and I don’t know what else.” If anything, this tendency has gotten even worse since.
Higher Education’s Hype Cycle
As that forum on AI exemplifies, it is easy to follow this ever-escalating hype cycle in the pages of The Chronicle of Higher Education. I wrote for The Chronicle for a few years in the 2010s, so I am loath to pick on it exclusively here. However, since The Chronicle has always published the “conventional wisdom” about higher education and offers readily available online archives, it is easy to compare the coverage of various technological “revolutions” at the time with what has actually happened since.
During the 1990s, the new technology was distance learning, what we now call online education. “Forget about arguments that it’s all just a lot of hype, these [campus] officials say,” exclaimed an article titled “The Coming Revolution” from 1994. “Forget about past disappointments over the unrealized payoffs of educational television, programmed learning, and videocassette recorders. This time there truly is a new revolution coming in American higher education—born of little 0’s and 1’s—and a lot of folks whose lives will be profoundly affected by that fact haven’t got the message yet.” Both the AAUP and the AFT noticed the difference in quality between in-person and online classes from the very beginning, and the persistence of that difference has prevented online education from destroying the on-campus educational experience throughout the country.
When smartphones started to appear in my classroom around the time that Apple released the iPhone in 2007, I remember looking at so many heads staring down at their phones and thinking that I must have become boring. Then I started talking to my colleagues and realized that I was far from alone. Yet, to The Chronicle, the smartphone “revolution” was administrative rather than pedagogical. “The phones,” explained one story—“originally a convenient way for students to talk or send messages to friends and family—have been grabbed by administrators and transformed into bulletin boards, classroom educational tools, and systems that alert campus security if a student doesn’t make it safely to her dorm at 3 a.m.” Based on my reading of The Chronicle’s archive, those now-cliché articles about whether to allow cell phones in class came only much later. By then, I had changed the way I lectured and introduced more interactive elements into all my classes, just as most of my colleagues did, and higher education went on largely as it had before.
During the 2010s, the new revolutionary technology was massive open online courses (MOOCs). In September 2012, at the absolute apex of MOOC hype, The Chronicle ran another forum, this one called “Can MOOCs Save Higher Education?” It also released an ebook, Beyond the MOOC Hype: A Guide to Higher Education’s High-Tech Disruption. Here’s a tiny piece of Chronicle stalwart Kevin Carey’s contribution to the forum: “Indeed, the future is so clearly one of universal access to free, high-quality, impeccably branded online courses that their presence can be simply assumed.” Needless to say, this prediction has not come to pass.
Such a long record of exaggeration should have given the commentariat pause when AI came along. Yet, in the forum that followed The Chronicle’s labeling AI “a clear inflection point in the history of higher education,” the only contributor who really seemed to take the editors’ premise to task was Emily Bender, the coauthor of the book The AI Con. Bender wrote that this technological fad wasn’t “any actual scientific breakthrough but rather the desperation of tech companies to try to recoup their massive investment in so-called AI.” By implication, university administrators (and presumably some higher education publications) have fallen for Silicon Valley’s misleading public-relations campaign hook, line, and sinker.
Adapting (Again) in the Classroom
The error behind the assumption that AI constitutes some kind of revolution has become more obvious as faculty experience with it has accumulated. “Now, well into the semester, it has come as a pleasant surprise to me and to many of my fellow teachers that the A.I. apocalypse that was expected to arrive in full force in higher education this fall, bringing an end to reading and writing and school as we know them, has not come to pass just yet,” wrote Carlo Rotella, an English professor at Boston College, in The New York Times last November. “Responding to the rise of A.I. by doubling down on the humanity of the humanities appears to be working, at least for the moment.”
Rotella outlines two kinds of responses that fall under his umbrella of “doubling down on the humanity of the humanities.” One is to turn back the clock and require that everything be done with pencil and paper. I respect this position because the people who adopt it obviously share the same goals as I do. However, I could never choose it myself. In history, it would require an overemphasis on rote memorization, which would be incredibly disruptive to the way I want to teach. Furthermore, given how much I depend upon cutting and pasting to polish my own writing, I recognize that it would significantly degrade the quality of my students’ writing, too.
I’ve gravitated toward what Rotella describes as the other major response to AI in the classroom, to “emphasize teaching the process of writing—breaking it down into a series of steps that a teacher can see and respond to—rather than simply grading the product.” This approach is much less disruptive (since I’ve been doing it all along, in a sense), and it has required only minor changes in the kinds of assignments that I’ve always created.
In general, I have started to make the requirements for every essay too specific for AI to fulfill. For example, if I require particularly important terms from my lectures to appear in midterm essays, the only way for students using AI to ensure that they appear is to enter all of them in a prompt. This produces long lists of unrelated historical terms in essays that don’t deserve good grades. By simply asking students to underline their thesis, I can make it harder for students who are using AI to write their papers, because computer-written essays won’t have any thesis to underline. Another tactic I’ve developed is to limit the allowable primary sources that students in my survey course can use for a research paper to a specific database or website. I have friends who require students to use Google Docs and provide the version history of the papers they write so that all of the work they did is visible.
Along with these kinds of changes, it makes sense to alter one’s expectations when grading. When my dean invited a guest speaker to campus last fall to talk about AI, she offered advice along similar lines that has stuck with me. She told my colleagues and me that the old “C” is the new “F.” In other words, while you might have been willing to pass a student simply for making an average effort in the past, you can no longer deem acceptable prose that a computer can create. This isn’t so much a change in pedagogy as a change in expectations that will serve students’ best interests in the long run. The term AI slop describes the weird, mash-up pictures and hyperrealistic but impossible cartoon videos that now appear on social media sites in order to capture the attention of your credulous older relatives. If the prose equivalent of those pictures becomes acceptable professional writing, every humanities professor in academia might as well just throw in the towel. That’s why I am one of those professors who has banned students from using AI in class for any reason.
Despite this position, I remain wary of failing any student for using AI because, unlike plagiarism, the evidence of AI use is never certain unless the student who used it confesses. (AI’s known proclivity to generate fake citations might also get them in trouble, but I have yet to encounter this situation.) Failure to meet specific assignment requirements can’t be disputed. While many of my colleagues believe that they can spot AI writing instantly, I remain less confident. I am, however, certain that AI produces bland, unspecific, pedestrian prose. With that slight adjustment in standards outlined above, I continue to grade that kind of work on its merits. Bad writing gets bad grades no matter who or what wrote it. It seems ridiculous to waste time using unreliable AI detection software when I can justifiably grade what’s in front of me with the same result.
The AI Bandwagon
With these strategies in place, the greatest nuisance that I associate with AI is not student cheating but administrative advocacy. The explosion of AI hype has led to administrators going much further than mine have to this point. To cite one large and particularly egregious example, the California State University system is attempting to turn over its educational decision-making apparatus to the AI industry. “Facing deficits and enrollment declines,” explained an article from Current Affairs about the AI initiative there, “administrators embraced the rhetoric of AI-innovation as if it were salvation.” The New York Times, citing Cal State as a prime example, went further, arguing that “some major universities are inviting tech companies, which typically supply campus computers and email, to take on a much bigger role as education thought partners, A.I. instructors and curriculum providers.” These kinds of programs elevate public-relations hype over the academic expertise of faculty.
The greatest potential threat to academic freedom posed by AI comes from precisely this kind of administrative encouragement. Ohio State University, another prominent proponent of AI, has begun an AI initiative designed to make every student “fluent” in this technology. “OSU leaders are adamant that they’re incentivizing and enabling AI use among the specific disciplines of staff and faculty—not mandating it,” reported Axios last year. Despite claims like this, the pressure on an untenured English professor to become an AI advocate would still be enormous. According to that same story, the entire initiative stems from the appointment of a new provost there, flying in the face of the AAUP’s guidance that the faculty should have primary responsibility for the curriculum.
Another important AAUP principle is that new technologies throughout the university should be adapted only with significant faculty input. Resisting administrative efforts to jump on the AI bandwagon may be the best use of shared governance available today as the hype surrounding this technology begins to evaporate. “The use of new technologies in teaching should be for the purpose of advancing the basic functions of colleges and universities to preserve, augment, and transmit knowledge,” explains the AAUP’s Statement on Online Education, “and to foster the abilities of students to learn.” Encouraging students to outsource critical thinking to AI is the exact opposite of fostering their ability to learn.
Of course, many faculty members—especially those outside the humanities—have a different opinion of AI from mine. I have no problem with that because academic freedom means that different faculty members can respond to different technologies in whatever way their academic expertise suggests. An English professor and an engineering professor will see AI differently because its arrival affects them and their disciplines in starkly different ways. Even if another history professor decides to embrace AI, I recognize that it is their prerogative to do so. While nobody in my administration has ever told me what I can or cannot teach with respect to artificial intelligence, I happen to think that the non-AI-related skills I teach my students will serve them better in the long run once the AI fad has run its course.
Despite my deep skepticism, I recognize that AI has some uses. One of my friends fed our notoriously opaque and contradictory faculty handbook into a program called Notebook LM in an effort to make heads or tails of the thing. I certainly don’t object to using AI for data collection or data processing (although I am concerned about its effects on the environment). The important consideration here is whether the faculty member using AI controls the way it gets used. In a situation where shared governance has disappeared, penny-pinching administrators will inevitably use it as a labor saving device, just as AI companies with sky-high valuations are telling their business customers to do. In this way, AI isn’t that much different from MOOCs, which required huge initial investments and have never generated significant returns, because the educational experience they provided never proved popular enough to become profitable.
As is the case with so many other aspects of American life, the sudden popularity of AI in higher education today comes down to money. Justin Raden, writing in Defector, connects such initiatives to the Trump administration’s decision to limit federal grants to universities for purposes that the president and his party do not support. “Born out of these now-distant federal research projects, then,” Raden writes, “AI returns to the university as part of a broad effort to further corporatize universities, vocationalize higher-education instruction, and diminish both research and research-based pedagogy.” If this somehow happens, AI will be so much more than just a nuisance. It will become an existential threat to the value of higher education as a whole.
Just because some tasks can be automated does not mean that they actually should be. I believe that the hype about AI changing everything in higher education is an attempt to intimidate faculty into making themselves obsolete by automating aspects of education that are best done by well-qualified human faculty members. If this attempt succeeds, wealthy students will continue to get a high-quality education, while poorer students will mostly get chatbots. I know that some professors are excited about using AI in their writing-based classes, but I don’t think teaching students how to give ChatGPT better prompts is a good use of anyone’s time, because I want to create a different future. AI’s obvious structural weaknesses have been obscured by its proponents’ rose-colored glasses.
Yet these proponents would have us redirect all of higher education in reaction to what may very well be a flash in the pan. Their bravado is a sign of the weakness of their chosen technology, not its strength. The ability of ChatGPT to write coherent sentences has hidden the obvious truth that the writing it produces is still very bad compared with that of educated human beings. AI can’t make judgments. It can’t carry an argument from the beginning of an essay to the end, because it can’t think. AI creates bad prose by giving form to the milquetoast median of nearly everything written on a subject that’s available on the open internet that can be captured in a large language model. Unless all that money going into its development can design a new technology from the ground up, AI’s writing skills are unlikely to improve. Unfortunately, this is no reason to become complacent. Even a blunt instrument can be used as a weapon if it falls into the wrong hands.
The struggle between administrators and the professoriate over control of the classroom will continue regardless of whatever new technology Silicon Valley creates. Therefore, ceding whatever prerogatives you have over your pedagogy and completely changing your teaching style because one technology has yet again “changed everything” is a terrible idea. AI is just one of a continuing series of technological nuisances that may merit some adjustments in teaching practices, but we can’t let it become a reason for higher education to drastically lower its standards. AI slop might find an audience on Facebook and Instagram, but students should never have to pay tuition to learn how to create it in a college classroom.
Jonathan Rees is professor of history at Colorado State University–Pueblo. He is copresident of the CSU Pueblo AAUP, a former copresident of the Colorado AAUP conference, and a former two-term member of the AAUP’s national Council. His email address is [email protected].