Why a Ban on AI in Classrooms Won’t Solve Poor Reading Skills in a Post-Literary Society

Policymakers are sounding the alarm about AI and social media in schools, as PISA’s latest report confirmed that student performance in reading continues to decline. French Education minister Édouard Geffray lamented the “cognitive destruction” of a generation, while his Spanish counterpart, Milagros Tolón, recommended a focused effort to improve reading because “reading is the foundation of everything”.

The Economist’s editorial board called it a “slow-moving catastrophe”, emphasising that “while machines are getting brighter (… ) teens are getting dimmer”, and pinned blame on the weak, “radical” leftists in university education departments. Meanwhile, the “radical left” is taking meaningful action. On September 2, shortly before the publication of the PISA scores, New York Mayor Zohran Mamdani announced a ban on student-facing generative Artificial Intelligence (AI) use in New York City public schools from pre-school through eighth grade (approx. 4-14 years of age), to be complemented by “twice-yearly AI critical thinking modules for high schoolers, limited AI pilots for a small number of high school classrooms and age-appropriate screen time restrictions”.

Mamdani’s policies are well intentioned. They offer a valuable symbol of resistance against a rapacious commodification of children’s attention, and they test a model which might be replicated beyond New York. Any measures which attempt to create space for young children to learn in an environment free of powerful and highly addictive AI technologies should be celebrated. But such policies address symptoms rather than confront the nature and causes of the decline in reading. This is because they arise from a superficial, reactive, and often patronising, public discourse, which has failed to honestly or accurately account for the scope of the transformation our society is experiencing: a technological shift from a literary to a post-literary society, in which AI is just one recent development, and through which all of our learning habits and capabilities are being remodelled, ministers and schoolchildren alike.

A media theory from the mid-twentieth century, developed by the classicist Eric Havelock and, perhaps most penetratingly, the Jesuit priest and literature professor, Walter J. Ong, can help us to design regulation in a way that is sensitive to the technological context in which AI has emerged: a context of transformation from a society which privileges one medium of knowledge production, communication and preservation, towards another. For the past hundred years, our society has shifted from a predominantly literary culture towards an increasingly oral and auditory one: from a culture which privileges interiority, reflection, and craft, towards one which values live communication, interactivity, and a certain ‘unfiltered’ and authentic immediacy. Understanding this context will help us to produce regulations which are more appropriately designed to attune institutions like the school to the technological context beyond their doors rather than attempt to hermetically seal them off.

AI is a machine for the de-sanctification of text. If we reconsider PISA in this light, the scores can be construed not as evidence of teenagers’ ‘dimness’ but of the acuity of their perception: perhaps these students have intuited that something has shifted irrevocably about the position of reading and writing in our culture. What if the students know something profound has changed about the world that they will live in that their teachers and ministers have either failed to understand or are unwilling to admit?

The foundation of everything

After a summer that welcomed a renewed interest in Homer, following the release of Christopher Nolan’s The Odyssey, it is worth remembering that the poems which form the foundation of the Western literary canon – the Iliad and the Odyssey – were composed, performed, transmitted, and preserved by people who could not read. As such, we might question whether reading ought to be considered the “foundation of everything”, as Minister Tolón suggests, or whether there is not a deeper substrate, on which reading itself depends, that we are missing in the controversy over our society’s changing relationship to the written word.

The classicist, Eric Havelock, in his 1963 Preface to Plato, argues that the birth of philosophy in Ancient Greece came about due to an innovation in the society’s methods for storing knowledge. Every human society relies on technologies to store information for reuse. Before Homer, Greek culture was stored in oral memory, but between Homer and Plato, the storage method changed with the invention of the technology of writing. The ascendancy of the written word brought about a crisis in Greek culture, which, according to Havelock, saw the replacement of an “orally memorised tradition” by a “quite different system of instruction and education”, and “saw the Homeric state of mind give way to the Platonic”.

Walter J. Ong, in his 1982 Orality and Literacy, describes how the gradual rise of this Platonic written culture, and its displacement of rhetoric (literally ‘speech arts’) from the centre of academic education, had a profound effect on the ‘mentality’ of learning in literate societies. Ong argued that writing reshaped consciousness by defining the ‘interior self’ against which “the objective world is set”. One product of this interiorisation, for Ong, was the emergence of the educational phenomenon of ‘study’ conceived as a reflective, solitary activity. Ong underlines that humans in oral cultures “learn a great deal and possess and practise great wisdom, but they do not ‘study’”. They learn relationally, he explains, by “discipleship, which is a kind of apprenticeship, by listening (…) not by study in the strict sense”.

Ong’s description of oral learning culture bears a similarity to the rhythms and immediacy through which knowledge is acquired, shared, and memetically elaborated on the Internet. Writing in the 1980s, Ong saw that the emergence of electronic technology (radio, television, and telephone) brought a new age of ‘secondary orality’. This new oral culture, which has evolved through Internet and digital technology, bore “striking resemblances to the old”. It unsettled the static quality of words, reemphasised dialogue and “agonistic” exchange, bringing a renewed “concentration on the present moment”.

The Gutenberg parenthesis

Throughout the period of literary culture – what Danish professor of literature Lars Ole Sauerberg has called the “Gutenberg Parenthesis”, a 500-year “blip” from the invention of print until roughly the turn of the millennium – text retained an almost sacred authority through its fixity and scarcity.

In the age of AI, text has become unfixed and maximally unscarce – chatbots produce unmanageably large quantities, and that text has come to resemble oral discourse. It is fluid and unreliable. On the Internet, dates and facts must be checked and triple checked. Spurious claims abound and must be interrogated.

Text, in literate societies, was “basically unresponsive”: if you ask someone to explain their position, Ong related, you get an explanation, but “if you ask a text, you get back nothing except the same, often stupid, words which called for your question in the first place”. In the new age of AI-mediated digital orality, the text explains itself. Or at least it appears to.

The PISA tests were designed in 2000 as the literary age – in which facility with text was the privileged sign of attainment – was reaching its apotheosis and giving way to the forces that would shape the new digital orality. The mind that could pay unbroken attention to a text, which contained a truth that would not change, and would not answer back, was the mind that our schools, our universities and our examination and assessment structures were built to produce and to measure. More fundamentally, this was the mind that our societies were organised to reward and incentivise. But even if we may wish to return the genie to the bottle, the world in which we live is thoroughly given over to a new logic of success and a new aesthetic of achievement.

Refusing the gift horse

Technology is not an enemy at the gates, but a gift horse that, bidden or unbidden, detected or undetected, is already within. It is already in the process of reshaping every major social institution.

Even if Mamdani’s policy keeps AI out of the physical classroom, the new reality that is imposed by the return to orality will continue to shape the incentives through which those classrooms function, by defining what kind of mind is valued and rewarded in the world outside.

It can take a long time for a society to understand that the technological context which made sense of its forms and rituals has changed.

In the Phaedrus, Socrates tells the story of the Egyptian king Thamus who refused the gift of writing because it would “create forgetfulness” and give “not truth, but only the semblance of truth”: students with writing, “will be hearers of many things and will have learned nothing”. But Thamus’s refusal – preserved in Plato’s written dialogue – was, like Mamdani’s, enveloped in the form of the medium it tried to refuse.

Mamdani’s moratorium may well be worth having. But the conversation around it should be less patronising, and the students themselves have to do more of the talking, because they may have grasped what their ministers have not: that we are testing them for a world which no longer exists. And if they are getting dimmer, we should have the humility to admit that we are getting dimmer too.

A weekly e-mail in English featuring expertise from scholars and researchers. It provides an introduction to the diversity of research coming out of the continent and considers some of the key issues facing European countries. Get the newsletter!