The librarian that never read by Dai Eun Greer

There is something almost darkly comic about a machine deciding which books are too upsetting for children, like a thermometer diagnosing grief or a lock choosing which doors deserve to exist. And yet, in a secondary school in Greater Manchester, this is not metaphor but policy, an AI system, fed a library catalogue, quietly disqualified nearly two hundred titles. Among the exiled were a former First Lady’s memoir, a wildly popular vampire romance and a graphic adaptation of a novel famous for warning us about systems that erase inconvenient truths. The irony, one suspects, was not flagged by the algorithm.

At first glance, the decision might appear practical. Schools are busy places; educators are overburdened; technology promises efficiency. Why not outsource the tedious work of cataloguing appropriateness to a system that can scan thousands of titles in seconds? But this is precisely the problem: appropriateness is not a measurable property like page count or publication date. It is a conversation, messy, contextual and deeply human.

When a machine rejects a book for containing “upsetting themes,” what it is really doing is flattening the entire purpose of literature. Stories are, by their nature, unsettling. They are meant to disturb complacency, to introduce discomfort, to complicate certainty. A young reader encountering difficult ideas, loss, injustice, moral ambiguity, is not being harmed; they are being initiated into the complexities of the world they already inhabit. To deny them that encounter is not protection. It is deprivation.

There is also a quieter danger in allowing AI to curate cultural spaces: the illusion of neutrality. Machines do not invent values; they inherit them. Somewhere, in the opaque layers of code or training data, a set of assumptions about what children should or should not see has been encoded. Those assumptions may be conservative, risk-averse or simply ill-informed. But because they emerge from a machine, they carry a veneer of objectivity that human judgment does not. A librarian can be questioned. An algorithm, too often, is obeyed.

This shift matters because libraries are not merely storage rooms for books; they are declarations of what a community believes is worth knowing. To quietly remove titles based on automated decisions is to abdicate that responsibility. It replaces a public, accountable process with a private, inscrutable one. The result is not just a thinner bookshelf but a thinner intellectual life.

There is, too, a pedagogical contradiction at play. Schools aim to teach critical thinking, yet here they model uncritical delegation. Students are told, implicitly, that difficult decisions, about what to read, what to question, what to engage with, can be outsourced to a machine. It is a lesson in passivity disguised as innovation.

None of this is to suggest that AI has no place in education. It can assist, augment, and even inspire. But it should not decide. The act of choosing books, especially for young minds, is an ethical exercise, not a technical one. It requires empathy, judgment, and an awareness of nuance that no system, however advanced, can replicate.

In the end, the image lingers, a silent program scanning titles, issuing verdicts, thinning a library without ever turning a page. It is efficient. It is scalable. And it is profoundly illiterate.