The Unspoken Cause (and Solutions) to Our AI Problems

Shortly after the end of the Spring 2026 semester, President Michael Elliott published an article in The Atlantic, “College Should Be Way More Fun,” on the value of the human “intellectual experience” amid the surge of students relying on artificial intelligence (AI) chatbots to complete their education nowadays. While I certainly agree with Elliott, I also wish to point out what he — along with many others — consistently overlook when talking about this important issue. Moreover, I wish to highlight the irony of Elliott’s critique given that institutions such as Amherst feed into the exact conditions that encourage students to rely on AI chatbots for their education.

First, I will briefly summarize the solution Elliott believes will remedy our AI woes. At the core of his argument is the concept of “intellectual play,” or inquiry which is inspired by curiosity and skepticism instead of righteousness and aversion to ambiguity, and how it should be the pedagogical emphasis of an undergraduate education.

Put more concretely, this means emphasizing dialogue and a tolerance for complexity in the classroom rather than relying on AI chatbots, whose ability to provide straightforward answers pales in comparison to the value of slow, messy conversation. Ultimately, in Elliott’s eyes, the result of this intellectual play would be a reintroduction of important “cognitive” habits back into public life, habits which are necessary for securing “democratic norms” or the building of “shared meaning across fractured communities.”

How could anyone disagree with Elliott here? Am I to say that dialogue in the classroom should be replaced by the mindless use of AI chatbots? Let me be clear here: The issue is not in the ideals that Elliott is proposing, but rather in his ignorance of the material forces (i.e., forces related to our capitalist mode of production) that have made our education so susceptible to the flood of AI chatbots.

This reorientation of attitudes, from idealizations to material analysis, also applies to AI more generally, which will also assist in making my following argument more clear. Kate Crawford’s “Atlas of AI” summarizes what it means to think of AI from a materialist lens in the line: “AI is neither artificial nor intelligent.” It is not intelligent in the sense that these AI systems require extensive training before being able to perform any so-called “autonomous functions,” and is not artificial in the sense that these systems are deeply embedded in the material world, both in the natural resources required to sustain them and the broader social structures they reflect and are embedded into. The latter part is the most relevant and important to this discussion since AI use by students is not merely driven by, to use the language of Elliott, an ignorance of the value of intellectual play or their cognitive habits. Rather, it is clearly a symptom of a much larger structural and systemic issue: the oppressive role education is reduced to under capitalism.

Paulo Freire, a Brazilian philosopher, educator, and one of the founders of modern critical pedagogy, spends the bulk of his work exploring how education takes up this oppressive role. Specifically, Freire uses the term “banking model” to describe how, under capitalism, education is reduced to an act of depositing, where the only task of the educator is to deposit information to their students who are the depositories. In this model, the scope of action allowed by the students is — like that of an automaton — restricted to patiently receiving, memorizing, and reciting the information bestowed upon them by the educator. Consequently, the students, seeing themselves only as ignorant, empty minds to be filled by the knowledge bestowed upon them by the teacher, adopt a passive and alienated existence. That is, they do not see themselves as having the ability to intervene in and transform the world, instead only to adapt to it.

Does this not sound all too familiar, both to students and educators? Let us brush aside the theory for a moment and look at the real, current state of our educational environment. Are we, as students, not familiar with the process of studying exhaustively for exams, putting our mental clarity on the line, only to then forget about the material immediately after? And are we, as educators, not currently dealing with a crisis of critical thinking amongst students as it relates to, among other negative trends, a decline of reading? Let it be absolutely clear: These are not defects of our educational model which have suddenly and randomly emerged. Rather, they serve as clear evidence of it working exactly as intended in the fashion of the banking model. Thus, it is not a coincidence that students’ use of AI chatbots fits so neatly into their educational habits. Receiving, memorizing, and reciting information are precisely the functions of AI chatbots, so when students are restricted to these same functions, they will obviously turn to and rely on such technology.

What, then, is the solution to our AI woes? Evidently, it is to ditch the oppressive banking-model and transition toward an alternative, liberatory model of education. Once again, Freire comes in handy here by providing us with that alternative model: dialogic education. In dialogic education (also called problem-posing education), learning is explicitly an act of cognition — not a transfer. Students and teachers cease to exist separately and are, instead, reconciled into student-teachers and teacher-students who jointly teach each other through dialogue. 

To illustrate what this looks like in practice, take Freire’s own example of the phrase “the capital of Pará is Belém.” Under the banking model, this is simply a phrase to be uttered by the teacher and then recorded, memorized, and repeated by their students. However, under the dialogic model, the students are invited to explore the true significance of the word “capital,” or what Belém means for Pará and what Pará means for Brazil. Consequently, the students here see themselves not as docile listeners, but rather as co-investigators in and with the world; who not only feel challenged by the problems of the world but also obliged to respond to that challenge.

I expect the reader to be confused. Is dialogic education not what Elliott was proposing with intellectual play? There is an important distinction to be had here: Education incorporating dialogue is not the same as dialogic education. The former is merely a classroom procedure while the latter is a comprehensive and revolutionary systematic project. That is, it not only applies to the structures of the classroom but also the structures of society at large, between leaders and the people. After all, an oppressive society will also be necessarily antidialogical. The oppressors, in order to retain their dominance over the oppressed, must convert them into passive, alienated spectators of the world around them.

Ultimately, this is the reason why institutions such as Amherst feed into the conditions which encourage students to rely on AI chatbots. While bluebook exams, harsh grade penalties, or classes on AI literacy could certainly mitigate the amount of students using AI, all these so-called “solutions” fail to address the root cause of the issue: the banking model of education which we follow. 

Again, this is not a defect of our approaches but proof that our system is working exactly as intended. While the college loves to preach those heartwarming ideals of critical thinking and democracy, it is evidently an oppressive structure in and of itself. 

First, consider the fact that any and all meaningful decisions related to the college must pass through a board of trustees whose interests are secured outside that of the faculty or students. Case and point, of course, was in 2024 when the college refused to divest from Israel despite overwhelming, near unanimous support from both the student body and the faculty. In denying them their wishes for divestment, the college demonstrates how democracy and critical thinking are respected only to the extent that it does not impinge on the flow of capital. Thus, in this domination over student and faculty interests by the elites on the board of trustees, there clearly emerges a distinct oppressive reality.

Second, consider the other institutions which Amherst has a deep relation with. I am talking specifically about the ties to high finance, corporate tech, big law, and other prestigious colleges and universities. Between all of these institutions, there is a common thread that I wish to highlight: They all require a supply of obedient workers who can fit neatly into a hierarchy and regurgitate the rigid patterns of anti-dialogic miseducation that they themselves were taught. For Freire, this explains how our current educational institutions constitute structures of oppression. Future generations, themselves victims of oppression, are transformed into its agents as they go on to reproduce the same miseducation they received, thereby reinforcing the broader structures of oppression. Thus, because educational institutions such as Amherst function largely to churn out these obedient workers who — above all else — reproduce their own miseducation, not only are they evidently oppressive structures, but they will certainly never not adhere to an anti-dialogic model of education.

So long as we persist with an educational model that treats knowledge as a gift bestowed upon the innately ignorant, we will never be able to solve the crisis of our students using AI chatbots to complete their education. There is a solution, as I have laid out, which is certainly not to emphasize intellectual play — as if that would provide any real, meaningful change. Rather, what we need is a complete transformation of how we structure our educational institutions; from reconciling the divide between student and teacher, to building an institutional democracy which places the interests of the student body and faculty first.