We are only at the beginning of the age of artificial intelligence (AI). Every powerful new idea arrives twice: first as a provocation, then as a problem.

It is not surprising to see that popular and scholarly responses have lurched between uncritical enthusiasm and reflexive alarm. But before we mourn what AI threatens, we should ask what we’ve been defending — and whether it deserves defending.

In Philosophy in a New Key, American thinker Susanne Langer argues that powerful ideas initially generate exaggerated hopes and fears before society settles into confronting the real problems they create. AI appears to be following precisely this trajectory.

Scholars and researchers are warning that overreliance on AI undermines independent thinking and erodes key cognitive and executive functions. Even literature professors are worried about AI. Micah Nathan, a fiction writing professor at the Massachusetts Institute of Technology (MIT), described feeling hurt upon discovering that two of his students had used AI to write their workshop stories. Meghan O’Rourke, a poet and creative writing professor at Yale, wrote of an unease she struggled to name — a sense that AI produces writing that is “mimetic of thought, but not quite thought itself.”

Critics worry that AI will erode the cognitive benefits of writing. But beneath those concerns lies a deeper assumption: that writing is the privileged medium of thought. History tells a more complicated story

However, the scholarly literature on AI and both professors’ essays share a common assumption that is so deep that it goes almost entirely unexamined. Their essays implicitly elevate writing as the highest and most authoritative form of thinking. That premise deserves scrutiny, especially at the beginning of the AI era.

Writing as an Instrument of Control

Scholars, including archaeologists, historians and anthropologists, have argued that writing did not emerge primarily as a vehicle for individual expression or cognitive development. Writing first developed as a tool of administration and accounting — not expression.

For instance, of all the written texts recovered from the early Mesopotamian kingdom of Uruk, 85 percent were economic records, such as tax receipts, population lists and cadastral maps of taxable land. As Belgian-French anthropologist Claude Lévi-Strauss observed: writing “seems rather to favour the exploitation than the enlightenment of mankind.”

The elementary form of statecraft, American political scientist and anthropologist James C. Scott argues in The Art of Not Being Governed, is the population census, which is the basis for taxation and conscription. The hill peoples of Southeast Asia called Akha, whom Scott studied, understood this perfectly well. When the Akha people recall their ancestral oppressor, the great crime they remember is not military conquest but a yearly census. Writing, in their experience, was not a tool of self-expression. It was the instrument through which they could be counted, taxed and controlled.

Historians argue that many of these people, such as the Akha, the Hmong, the Karen and the Lahu, were not pre-literate; rather, they were post-literate. They had known writing, been close to literate lowland states for 2,000 years and, in many cases, actively moved away from it as a survival strategy.

An oral culture, Scott shows, is a “jellyfish” culture, meaning it is constantly shapeshifting, adaptive, resistant to being fixed and administered. This is a deliberate choice. In contrast to orality, writing freezes; it makes you legible to power. Orality keeps you mobile.

This is not to romanticise oral cultures, which have their own hierarchies and constraints, but to recognise that literacy has never been a politically neutral technology.

A tradition mistaken for universal

AI, ironically, may be doing something similar to the written word itself — revealing who the essay form was really built for.

When Nathan writes in his essay that “writing is both vehicle and vessel for thinking — abstract made concrete, feelings translated into words”, he is describing something real about his own tradition. The Western philosophical canon, such as Plato, Montaigne, Descartes and Orwell, built its identity around written thought, around the essay and the argument and the revision. O’Rourke puts it even more plainly: “…language is our most human inheritance: the space of richly articulated perception, where thought and emotion meet.”

These are beautifully composed sentences. But they are sentences written from within a specific civilisational tradition, and they universalise that tradition without acknowledging it. The griot of West Africa, who preserved genealogies and histories in memory, and the Akha phima of Southeast Asia, who recited law and lineage at ceremonial gatherings, performed the same cognitive labour Nathan and O’Rourke associate with writing, only through oral forms.

The assumption that writing occupies a privileged place in human cognition is not a neutral observation about how people think. It is a civilisational claim, one that, historically, has been used to stigmatise non-literate peoples as “barbarians without history.”

What writing has done to us

The deeper problem with the two literary professors’ essays is that they ask what AI does to writing, but not what writing has done — historically, structurally, in terms of who it has included and excluded, what kinds of minds it has rewarded and which it has quietly failed. The workshop that Nathan defends is a powerful institution. Critics have long argued that it produces a relatively homogeneous literature, shaped by the tastes and tolerances of those who could afford to sit in its rooms and suffer productively.

This is not incidental. Written literary traditions, including the essay, the workshop story, the lyric poem, have always carried within them a set of implicit cognitive and cultural requirements. For instance, a relationship to solitude, linear argument, and the belief that one’s inner life is worth recording. The griot did not need these things. The phima did not need these things. They had other requirements, other disciplines, other ways of making thought audible and transmissible.

What they did not have was a tenure committee, a literary journal or a workshop. If literacy has historically been bound up with institutions of administration and authority, it is worth asking whether modern literary institutions carry their own forms of gatekeeping.

When O’Rourke writes that language is “our most human inheritance”, the possessive is doing a great deal of unacknowledged work. Whose inheritance, exactly? The Hmong oral poet’s? The Lahu elder’s? The first-generation college student who thinks in one language and is required to write in another, whose cognitive life has never mapped cleanly on to the essay form, and who has been told, implicitly and explicitly, that this gap is a deficiency to be corrected rather than a difference to be reckoned with?

The workshop has not been neutral about these questions. It has had answers, mostly unspoken, about what good writing sounds like, what a reliable narrative voice is, what counts as precision and what counts as excess. Those answers have not been universal. They have been particular — specific to a tradition, a class, a language, a set of institutions that developed in specific historical circumstances and then forgot that they had.

None of this means writing is merely an instrument of domination. Writing has enabled forms of reflection, preservation and critique that would be impossible otherwise. If AI disrupts that institution, the disruption is not only loss. It may also be, for some people, a kind of freedom. The question worth sitting with is not whether AI threatens the culture of the written word. It clearly does, in some ways.

The question is: which parts of that culture are worth defending, and which parts have we been defending — without quite realising it — because they were always already doing the work of keeping certain people out.

Farah Adeed is a PhD Candidate in Political Science at Boston University in the US

Dr Muhammad Zubair Abbasi holds a DPhil in law from the University of Oxford. He is based at the School of Law at Royal Holloway, University of London in the UK

Published in Dawn, EOS, June 21st, 2026