I began teaching at Towson University, a medium-sized public institution in Maryland, in January 2002. Even then, it was a challenge to develop assessment tools that genuinely fostered critical thinking. The No Child Left Behind era rewarded rote memorization and compliance—what do we have to know to pass?—rather than knowing how to think. Under this curriculum, complexity and nuance were not valued. I say this to avoid slipping into “old lady yelling at cloud” nostalgia for an educational system that did not reliably foster or reward critical thinking for a generation.
Many students today attend college to earn a marketable degree. Higher education functions more as a very expensive, debt-driven job-training pipeline than a site for learning to think critically about the world. Employers benefit enormously from outsourcing job training to colleges and universities, and it serves their interests to have desperate employees buried in debt and therefore subservient to the needs of capital. These mounting costs push students to complete degrees as quickly and “efficiently” as possible. Many struggle to get the work done at all, let alone to do it well.
And let’s be honest: Students have always cheated. They paid others to write papers. With the rise of search engines and Wikipedia, they stitched together hodgepodges of text lifted straight off the web. But no technology has disrupted education as quickly or as profoundly as generative AI.
Those Were the Days?
Again, it is tempting to imagine a golden age when learning for its own sake was universally valued. Over a century ago, Thorstein Veblen warned that university administrators had begun acting more like business executives than stewards of higher learning by turning education into a marketable product, valuing scholarship for its profit potential, expanding managerial control, and treating faculty work as replaceable labor. In his view, this shift pushed universities away from intellectual life and toward financial metrics.
I attended college at the height of the 1990s neoliberal restructuring that intensified the instrumentalization of higher education. Sheila Slaughter and Gary Rhoades have identified universities during this period as operating within a regime of “academic capitalism,” in which students are revenue streams and products to be crafted and instruction is workforce preparation; personal or social transformation is secondary at best. All of this is to say that higher education has long been entangled with capital. AI has simply supercharged that entanglement.
That said, when generative AI first appeared, administrators seemed as alarmed as faculty members. It seemed to come out of nowhere. Many administrations tacitly encouraged faculty to ban AI while we tried to understand what we were facing. That sentiment quickly passed. Now our institutions openly embrace AI and are seeking to incorporate it into “learning” at every level.
Across the university, AI is being built into almost every technology. Rather than heeding the many warnings about AI—environmental, political, social—university leaders are uncritically heralding AI as a vehicle for preparing students for the workforce of the future. Their rhetoric reinforces the mandate that college is more a means to build credentials than a place where intellectual development is encouraged and cultivated.
My own university masks AI use as anticorporate empowerment. The authors of a draft report from TU’s Generative Artificial Intelligence Task Force state, “Rather than let this conversation be decided by external forces (corporations and politicians), universities need to step up and contribute to debates about something that threatens (and promises) to change the basis for university work.” This stance is understandable if you take the spread of AI through higher education as inevitable. And maybe it is. But the sentiment of inevitability contributes to an acceleration of a process that has long threatened university learning.
The university has for generations been on a trajectory that places ever-greater emphasis on efficiency and market alignment. Generative AI fits seamlessly into this logic. According to Ronald Purser, in his excellent Current Affairs article, “AI Is Destroying the University and Learning Itself,” “AI centers are flourishing and institutional resources are being diverted into their distribution. At the same time, humanities and social science departments are being shuttered due to declining enrollments and ‘lack of financial viability.’” These developments align with the neoliberal logic of the university, wherein the goal of higher education is to produce more competitive job candidates, especially in a shrinking labor market in which AI itself is a key driver of displacement. AI accelerates the demise of a system already fighting for its life.
Outsourcing Cognitive Labor
Short of buying a paper or hiring someone to write it, students engaged in earlier forms of cheating still had to expend some mental labor. Even pasting together results found on Google at least made the student engage with the ideas and cobble them together, however haphazardly. I have taught a mandatory first-year writing seminar for nearly a decade. Before AI, plagiarism was easier to identify—copied abstracts, unattributed quotations from Wikipedia (at worst) or journal articles (at best). When students were unaware that their actions constituted plagiarism and made a good-faith effort to correct it, we could treat it as a learning opportunity. Even if the final paper was weak, students were forced to wrestle with the discomfort of writing. AI has changed this completely.
For instance, in a New York Magazine piece titled “Everyone Is Cheating Their Way Through College,” a frustrated music professor reported that almost all of the students in one class turned in versions of the same essay. The article quotes him as stating, “They’re using AI because it’s a simple solution and it’s an easy way for them not to put in time writing essays. And I get it, because I hated writing essays when I was in school. . . . But now, whenever they encounter a little bit of difficulty, instead of fighting their way through that and growing from it, they retreat to something that makes it a lot easier for them.”
As a beloved professor once told me, writing is thinking. Even students who wrote poorly constructed papers still had to reason their way through them. Historian Steven Mintz captures this struggle well in his analysis of AI and the other forces dismantling higher education: “The traditional forms of evidence for learning—original writing, independent analysis, sustained engagement with complex texts—have been rendered unreliable. But the challenge runs deeper than assessment. AI has created an epistemic crisis for students themselves. Why struggle to formulate an argument when ChatGPT produces one instantly? Why spend hours wrestling with a difficult text when an AI can summarize it in seconds?”
AI is, above all, the most efficient tool yet developed for bypassing the learning process. In his Current Affairs article, Purser makes a critical distinction between a tool and a technology. Citing philosopher Peter Hershock, he observes that we use tools to accomplish a task, retaining some amount of agency and creativity in the process. A technology, in contrast, fundamentally changes the conditions and context of the process—in this case, learning. Purser highlights the distinction this way: “A pen extends communication without redefining it; social media transformed what we mean by privacy, friendship, even truth.”
AI companies routinely talk out of both sides of their mouths when it comes to taking a position on this issue. On the one hand, they minimize their significance, frequently describing AI as a “tool” to help people maximize productivity. On the other, they acknowledge that the technology was fundamentally created to change the cognitive landscape as we know it. Consider this manifesto from Cluely, an AI start-up:
We want to cheat on everything.
Yep, you heard that right.
Sales calls. Meetings. Negotiations.
If there’s a faster way to win—we’ll take it.
We built Cluely so you never have to think alone again. . . .
AI isn’t just another tool—
It will redefine how our world works. . . .
The future won’t reward effort. It’ll reward leverage.
So, start cheating. Because when everyone does, no one is.
AI companies and universities alike are leaning into the inevitability of an AI-driven cognitive landscape. And the preliminary research on the effects is damning.
Cognitive Debt
If it were simply a matter of a few students cutting corners, the problem would be manageable. But it isn’t. Students increasingly report that AI use feels ubiquitous and compulsory. Integrating AI into teaching and learning is now framed as “staying competitive” by institutions and students alike. As Cluely says, if everyone is doing it, it’s no longer cheating. In a culture shaped by performance metrics, grade anxiety, and credential scarcity, opting out is beginning to feel like self-sabotage. My students seem as alarmed and dismayed by these technologies as I am. But anecdotal evidence (the number of papers that my colleagues and I receive that have clearly been written by AI) and early research indicate, in fact, that use of and demand for AI among university students is ever-growing. A 2025 national survey from the Pew Research Center indicates that the number of young people using ChatGPT to complete schoolwork has doubled since 2023. More tellingly, a study conducted by the Digital Education Council shows that a majority of students want and increasingly expect AI to be integrated into their education.
This sense that “everyone is using it” goes a long way toward explaining AI’s rapid spread, even among students who are well aware of its shortcomings. Social-norms research consistently shows that people tend to match their behavior to what they think others are doing. So, once AI use feels normal, or gets exaggerated through rumor, the psychological barrier to opting in drops fast. And that’s where the collective action problem kicks in: Even students who genuinely care about learning start to feel like they have to use AI just to keep up.
Students are turning to AI not to think with them but for them. What they might not understand is that the process of protracted effort is the point of doing academic work. Research shows that struggle and intellectual friction are essential to cognitive development. For example, a key study of the effects of large language model (LLM) use, “Your Brain on ChatGPT,” found that students who write essays with AI assistance exhibit significantly weaker neural connectivity than those who do not. Over time, students exhibit reduced cognitive engagement even when they aren’t using AI. The study also found that students who consistently rely on AI exhibit the lowest sense of ownership over their writing and struggle to recall their own work accurately. Most important, it found that long-term use of AI results in what the authors call “cognitive debt”—“a condition in which repeated reliance on external systems like LLMs replaces the effortful cognitive processes required for independent thinking.”
When students rely on the cognitive shortcuts AI provides, effort in the moment is reduced, which feels lifesaving for overstretched students forced to balance heavy course loads and long work hours while also hoping to have a personal life. Over time, however, long-term costs accumulate (think about the impacts on attention span attributed to social media use, for instance). The “cognitive offloading” on which students rely to get through their busy lives, or simply to avoid the frustration of the writing process, consistently results in a compounded and cumulatively diminished capacity for critical thinking.
AI reduces the friction of learning. What many students fail to understand is that friction is the point. The difficulty of synthesizing, organizing, and articulating ideas is the mechanism by which learning happens, and students are all but encouraged to bypass that process by our culture and, increasingly, by the university itself. The consequences extend far beyond the individual.
Education as the Practice of Freedom
The cognitive implications of the AI industrial complex are increasingly clear, but the political implications may be even more alarming. Paulo Freire famously distinguished between a banking model of education, where students are treated as passive consumers, and a problem-posing model, where learners become cocreators of knowledge, granting them agency in the learning process. AI supercharges the banking tendencies Freire warned against. It deposits prepackaged arguments and simulates dialogue without the friction of genuine analysis that real learning requires. In this sense, AI risks becoming the ultimate banking apparatus, accelerating the regressive banking model and moving us further away from critical and reflexive pedagogy. The timing couldn’t be worse, as it is coupled with unprecedented attacks on higher education and democracy itself, with programs that cultivate critical thought squarely under the hatchet.
There is a reason conservatives target the university. They are not entirely wrong that higher education liberalizes—and sometimes radicalizes—people. That was certainly my experience, and the data back it up. A 2016 study from the Pew Research Center found that college graduates are more likely to lean left than those without degrees. Citing the widespread adoption of diversity, equity, and inclusion programs since 2020, the Right has spread a moral panic about students “catching the woke mind virus.” This fear is being mobilized to dismantle the humanities and social sciences—the fields most likely to adopt critical pedagogical practices. At the same time, universities have been widely reframed as debt-driven career centers that offer diminishing economic returns. This confluence of trends further diminishes the role of the university as an incubator of critical inquiry. The rise of AI intersects with these trends in politically perilous ways.
Literacy, independent analysis, sustained attention, and tolerance for ambiguity are precisely the capacities that erode in an AI- and algorithm-driven landscape. Democratic life requires a polity capable of independent and nuanced judgment. A citizenry that increasingly outsources thinking to algorithms becomes vulnerable to manipulation and demagoguery. When sustained attention atrophies, so does the ability to track complex arguments or interrogate political claims. As Purser puts it, “Critical pedagogy is out; productivity hacks are in.” In a political climate hostile to intellectual autonomy (despite claims that people should “do their own research”), this shift is a godsend to the conservatives who rely on a pliant and uneducated populace.
What We Do in the Classroom
At the intersection of these trends, educators are scrambling to figure out how to salvage the vestiges of what feels like a dying system. Mintz, writing on Substack, distills the trajectory of the trends beautifully: “Ubiquitous technology fractured students’ attention. Economic pressure instrumentalized learning. Therapeutic culture reshaped students’ relationship to intellectual challenge. And now artificial intelligence has rendered our primary tools for assessing learning—essays, exams, analytical writing of any kind produced outside our direct observation—functionally obsolete.”
In response, many faculty members, me included, are shifting to in-class writing, handwritten annotations, group work, and collective reflection. Some colleagues have stopped assigning papers altogether, exhausted by reading AI-generated slop. Something fundamental is being lost, and we have no guidance from administrators, many of whom eagerly promote AI, about how to preserve it.
Writing a paper, doing close readings, and engaging in complex problem-solving—these are the very processes through which thinking happens. They cannot be replaced by off-the-cuff reflections or quick in-class exercises. The real question is whether universities will organize themselves around sustaining the cognitive and relational labor that a democratic society requires or will drift toward a model in which thinking itself becomes optional. If writing is thinking and thinking is labor, then the work of education cannot be automated without consequence.
In my view, the question is not whether AI can be used strategically. The question is what kind of minds, and by extension what kind of society, AI will cultivate. If frictionless cognition becomes the norm, if outsourcing the very labor that forms judgment is normalized, then the logical outcome is the erosion of independent thought and increasingly fragile democratic institutions. Universities were never pure sanctuaries of intellectual freedom, to be sure. But they have been one of the few institutional spaces where sustained attention, argumentation, and the slow work of thinking were cultivated and practiced. Automating intellectual labor in the name of efficiency and market alignment risks producing graduates who are ostensibly credentialed but cognitively hollowed out. As philosophy professor Martha Nussbaum has discussed, education cultivates human development in many ways, especially the “critical thinking [that] is particularly crucial for good citizenship in a society.” The price of outsourcing the university’s work to AI will be paid not only in diminished critical thinking in the university but also in a diminishment of civic life itself.
Heather Hax is assistant teaching professor in the Department of Sociology, Anthropology, and Criminal Justice at Towson University and a member of the AAUP’s Committee on Contingency and the Profession.