There is something unsettling about the idea of a book being bought, cut apart, scanned and then discarded, not because its contents are considered dangerous, false or culturally worthless, but because the physical object has ceased to have any further economic or technological use once its information has been extracted. The contemporary practice of destructive digitisation, brought into sharper focus by reports of artificial intelligence companies acquiring large quantities of books (most recently by The New Yorker), removing their bindings, scanning their pages and disposing of the physical copies, therefore raises a question that is considerably larger than the ethics of data acquisition. It asks what happens when we stop understanding an object as something in itself and begin to understand it only in terms of what can be extracted from it.
The destruction of books for artificial intelligence training represents the culmination of a much older philosophical movement in which the world is progressively transformed from a collection of things with histories, meanings and intrinsic qualities into a collection of resources available for human use. What is disturbing about destructive digitisation is therefore not simply that books are being destroyed, but that the destruction reveals a particular conception of knowledge in which information is separated from material existence, utility is separated from meaning, and the object survives only for as long as its contents remain economically or technologically valuable. Every book pulled into an AI training pipeline is, in this sense, a small case study in what happens when material history is treated as an obstacle to data.
This is not an argument that artificial intelligence companies are equivalent to historical book burners, nor that digitisation itself is destructive. The historical destruction of libraries was often an attempt to erase memory, suppress identity or prevent knowledge from surviving. Destructive scanning is almost the opposite. It destroys the object because it wants the knowledge inside it. Yet that distinction makes the phenomenon philosophically interesting rather than less so, because in both cases an authority makes a decision about what part of a book needs to survive and what part can be sacrificed.
The question is therefore not simply whether the information survives. It is whether information is all that a book is.
Long before anyone imagined a machine capable of ingesting millions of books, Sigmund Freud had already given Western thought a language for thinking about the strange relationship between creation and destruction. In Beyond the Pleasure Principle, published in 1920, Freud proposed the controversial concept of the death drive, an impulse towards repetition, dissolution and a return to an earlier, less complex state. It would be excessive to apply Freud directly to artificial intelligence companies, and there is no serious basis for suggesting that the engineers involved in destructive digitisation are psychologically motivated by a literal death drive. What Freud gives us instead is a way of recognising that human beings can become fascinated by processes of reduction and destruction even when those processes are presented as rational, productive or necessary.
A book is made through accumulation. Words accumulate into sentences, sentences into chapters, chapters into a work, and the work acquires a physical existence through paper, ink, binding, ownership and use. Destructive scanning reverses that accumulation. The book is reduced to pages, the pages to images, the images to data, and the physical object is then treated as redundant. What has been created digitally is preserved, while what made that digital representation possible is eliminated. Freud's fascination was with the individual psyche; what we are watching now is the same reductive impulse operating at the scale of an entire AI training pipeline for books.
A book contains more than its text
It is here that Martin Heidegger becomes particularly important, and arguably the single most useful philosopher for making sense of artificial intelligence's effect on the physical world. Heidegger's philosophical concern with technology was not simply that machines were becoming more powerful. His deeper concern was that technology represented a particular way of revealing the world. In his later writings, particularly The Question Concerning Technology, he described this technological mode of seeing as Gestell, usually translated as "enframing". Under enframing, things appear primarily as resources available to be ordered, measured, extracted and used. A forest becomes timber. A river becomes a source of hydroelectric power. The landscape becomes something to be surveyed and exploited. A library, under the same logic, becomes an AI training set.
The crucial point is that the danger lies not in the individual machine but in the way of seeing that accompanies it. Heidegger's standing reserve describes a world in which things are encountered primarily according to their potential availability for use. Technology does not merely give us new tools with which to interact with reality. It changes what reality looks like to us. His concern was that the technological mode of revealing could become so dominant that other ways of encountering things become increasingly difficult to recognise.
An AI training pipeline, in which every book becomes available on demand as data, is arguably standing reserve in its purest form: nothing withheld, nothing kept for its own sake, everything on call.
The book is a remarkably good object through which to understand this problem because a book can be many things at once. It can be a container of information, certainly, but it can also be an artefact, a gift, a possession, a record of ownership, a historical witness, a teaching object, a source of marginalia, a material expression of a particular period and a physical survivor of events that occurred long before the person holding it was born. The technological gaze has a tendency to flatten these distinctions. It asks instead: What can be extracted from this?
That question brings us to Walter Benjamin, whose 1930s writing on mechanical reproduction provides perhaps the most direct philosophical bridge between the physical book and its digital afterlife. Benjamin's famous discussion of the "aura" of the original work of art was not simply an argument that originals are expensive or sentimental while copies are cheap. He was concerned with what happens to an object's relationship to history when it becomes endlessly reproducible.
For Benjamin, authenticity was connected to the physical duration of an object and to the historical testimony contained in its existence. An original bears traces of where it has been, who has owned it and what has happened to it. Reproduction can preserve an image or representation while separating that representation from the object's particular history. What Benjamin is describing here is, in effect, provenance: the traceable chain of ownership and use that gives an object its authenticity, whether that object is a painting, a manuscript, or a parish register.
This becomes extraordinarily relevant to books.
Imagine two copies of exactly the same edition of a novel. The text may be identical, but one copy may have belonged to a soldier who carried it through a war, another may contain annotations made by a famous scholar, another may have been purchased from a particular bookseller in 1923, and another may contain a pressed flower, a handwritten dedication or an inscription that connects it to a family history. None of those things necessarily matters to a machine learning model attempting to understand the linguistic patterns contained within the text.
They may matter enormously to a historian.
Benjamin therefore gives us a distinction between content and testimony. A scan can preserve what the book says while destroying evidence of what happened to the book. The information survives, but the object loses its ability to testify through its physical existence. This is why the phrase "the information has been preserved" is not quite sufficient. It has been preserved in one sense, but something else has been lost.
That problem becomes even clearer when we move to Clement Greenberg, whose theory of medium specificity developed during the twentieth century. Greenberg argued that artistic practices should confront the particular properties of their own mediums rather than pretending that the medium is merely a transparent vehicle for something else. Painting, for example, has its own material conditions, and those conditions are not simply obstacles standing between the viewer and an abstract idea. Although Greenberg was writing about modernist art rather than books or artificial intelligence, his argument offers an important way of thinking about the physicality of the book. A book is not simply text temporarily stored inside paper. Its material form determines aspects of how the text exists in the world. The size of the page, the typography, the binding, the quality of the paper, the relationship between verso and recto, the marks made by readers and the physical arrangement of the object all contribute to its identity. The mistake is to imagine that the medium is merely packaging. The book's material form is part of its history.
Hannah Arendt takes this question somewhere even deeper. In The Human Condition, published in 1958, Arendt distinguishes between labour, work and action, and associates work with what she calls "worldliness". Human beings create a durable world of objects that outlasts individual human lives. Buildings, tools, works of art, books and other artefacts form a human made environment that gives continuity to generations. This idea matters enormously in an age in which physical objects are increasingly regarded as inefficient versions of their digital equivalents. The point of a book, from an Arendtian perspective, is not exhausted by the information contained within it. The physical book belongs to the world we inherit and subsequently leave behind. It is one of the ways human beings create continuity between people who will never meet. A book printed in 1890 can sit on a shelf in 2026. Someone who died more than a century ago can have touched it. Someone else can annotate it today. A future researcher can discover something in it that nobody at the moment of digitisation thought worth recording. Its continued existence allows future questions to be asked of it. Destruction closes those possibilities. This, in essence, is what an archive actually is: not a container for extractable information but part of the durable world Arendt describes, which is exactly why so much of it still resists being fully absorbed into a database.
This is where Erich Fromm introduces another dimension, because his distinction between biophilia and necrophilia was concerned with fundamentally different orientations towards life. Fromm used "necrophilia" in a philosophical and psychological sense to describe an attraction towards what is dead, mechanical, rigid and destructive, contrasting it with an orientation towards life, growth and creation. His discussion of what he called a "machine minded" culture is particularly suggestive here, although it would be absurd to describe the destruction of books as evidence of literal necrophilia.
Fromm's relevance is instead found in the question of what kind of world we construct when efficiency becomes the overriding measure of value. The living thing is complicated. It grows unpredictably. It accumulates history. It resists perfect standardisation. The machine prefers repeatability, classification and control.
It is the same undercurrent now surfacing, in far more alarmed terms, in this week's warnings from Jacob Coxon, the Anthropic and OpenAI researcher who resigned citing the existential risk of a purely mechanical logic of progress left unchecked.
The old book is inconvenient in precisely this way.
It has edges, imperfections, annotations, smells, stains, provenance and accidents. It cannot be completely reduced to the clean informational abstraction that a computational system prefers. Its imperfections are not necessarily defects. They may be evidence.
When preservation becomes extraction
For much of the history of library digitisation, copying has been associated with preservation and access. Fragile manuscripts were photographed so that researchers could consult them without repeatedly handling the originals, deteriorating newspapers were microfilmed before their paper failed, and digital collections allowed a researcher in London to examine material held thousands of miles away without threatening the survival of the object itself.
The underlying relationship was straightforward: reproduction allowed the information to travel while preservation allowed the source to remain.
Destructive scanning reverses that relationship, because the copy is no longer created to protect the original; the original is physically dismantled in order to create the copy.
This is a much more important reversal than it initially appears, because it represents a shift from preservation to extraction, in which the purpose is no longer to make a source accessible while safeguarding it, but to obtain the component of the source that has been identified as useful.
The book is gradually translated into another vocabulary. Its pages become images, the images become text, the text becomes tokens, the tokens become training data, and the training data contributes to a model, while everything that cannot participate in that conversion risks being treated as residue. It is exactly this gap, between what gets extracted and what gets left behind, that archive research as a discipline exists to work inside.
This is the thinking behind our Source campaign, Get Off My Back, which refers literally to the backs of books being cut away but also to the broader idea that materiality itself has begun to be treated as an inconvenience standing between technology and the information it wishes to extract.
When the Library Becomes a Dataset: Heidegger's Theory of Enframing
Martin Heidegger's philosophy of technology feels unexpectedly relevant here because his concern was never simply that machines might become more powerful, but that technological thinking might change the categories through which we understand the world.
He described a process in which things increasingly reveal themselves as resources awaiting ordering and use, so that a forest is encountered as timber, a river as energy and land as mineral potential.
The contemporary equivalent is remarkably easy to imagine: a library becomes data.
Once the book has been conceptually transformed into latent training material, the physical features that previously constituted the book begin to look inefficient. The binding obstructs the scanner, the spine slows the process, page turning consumes time, and the object's integrity becomes friction within a system designed around extraction.
At that point, cutting the spine can stop looking like destruction and begin looking like optimisation, which is precisely why the language we use matters so much.
“Destructive scanning” has a clinical neutrality that obscures an uncomplicated physical reality: when the process is complete, that particular book no longer exists.
The fire and the scanner
The history of book destruction makes this transformation especially uncomfortable, although the comparison requires care because the motives behind historical library destruction and artificial intelligence data acquisition are profoundly different.
The libraries of Warsaw, Jaffna, Sarajevo and Mosul were destroyed in circumstances involving war, ethnic violence, political domination and ideological persecution, and it would be both historically and morally wrong to suggest that a corporate scanning operation belongs in the same category.
The comparison becomes interesting precisely because the intentions are so different.
The historical biblioclast often destroyed books because the ideas, identities or memories contained within them were considered dangerous and therefore had to disappear. The artificial intelligence company wants the contents of books for almost the opposite reason, because professionally produced human writing is extraordinarily valuable material from which language models can learn.
Anthropic's own search for books, unsealed in court documents, reflected their value as complex, organised, long form human text.
The traditional biblioclast destroys the book because the knowledge must not survive, whereas what I would call the extractive biblioclast destroys the book after ensuring that the knowledge has been captured.
One seeks erasure while the other seeks ingestion, but both require a decision about which aspect of the book is allowed to continue into the future.
The historical destroyer concludes that neither the book nor its knowledge should survive. The extractive system decides that the informational content should survive while the material source may not.
The fire and the scanner therefore operate in opposite directions, yet they lead us towards the same philosophical question about who has the authority to determine the form in which human memory deserves to endure.
Reproduction without preservation
Walter Benjamin famously examined what happens to cultural objects when technologies of reproduction separate them from their original material circumstances, although the printed book complicates his idea of the unique original because books have always existed as multiples.
What feels particularly strange about destructive AI scanning is that it produces something more radical: reproduction that depends upon the destruction of the object being reproduced.
For centuries, one of the great advantages of reproduction was redundancy. Copies protected knowledge from the vulnerability of a single object, which is precisely why texts could survive fires, wars, accidents and the deterioration of individual volumes.
We have now arrived at a situation in which the creation of the digital copy can itself become the reason that the physical copy is destroyed.
The technology that should make preservation easier has, under a different economic logic, made destruction efficient.
From circulation to enclosure
There is another loss that is harder to quantify because books do not simply exist; they circulate, and through circulation they acquire biographies.
Someone buys a book, reads it, writes inside it and eventually sells it. Another person finds it twenty years later, carries it somewhere else, places something between its pages and eventually gives it to a library. Its price changes, its associations accumulate and its historical meaning develops through encounters nobody could have predicted when it left the printer.
A book acquires provenance because it remains available for history to happen to it.
A copy purchased for destructive scanning enters a terminal transaction because there is no next reader of that physical object, no later bookseller, no future donation and no researcher discovering it unexpectedly on a shelf fifty years from now.
Its language may continue to generate enormous value inside a proprietary computational system while the biography of the object from which that language was extracted has permanently ended.
This is why the concept of enclosure is useful, because knowledge that once circulated through a physical object between human beings is acquired, extracted and incorporated into private computational infrastructure, after which its informational value can continue to expand while the source itself has disappeared from circulation.
AI needs books because books are human
Perhaps the greatest irony is that artificial intelligence wants books precisely because they represent such concentrated expressions of human intelligence.
Books contain the accumulated labour of writers who spent years constructing ideas, historians who searched archives, scholars who verified evidence, editors who refined arguments, translators who moved thought between languages, publishers who exercised judgement, librarians who catalogued collections and readers who preserved particular copies long enough for them to reach us.
That human density is precisely what makes books so attractive as AI training data.
We have spoken casually for years about “feeding” information into artificial intelligence, but destructive scanning gives the metaphor an uncomfortable physical reality because the machine is being taught to write by consuming the objects through which generations of human beings taught one another how to think, while the process of consumption can involve dismantling the very artefacts from which that knowledge was obtained.
This does not mean that every book should be preserved forever, because libraries must deaccession material, publishers pulp stock, collections change and countless genuinely interchangeable copies exist.
The argument is neither sentimental nor anti technology.
It is about intellectual humility.
We should be cautious about assuming that because we have extracted the information we currently recognise as valuable, we have exhausted everything a physical source might ever be able to tell us. That caution is precisely why vetted, in-person archive research still matters, even in an age of AI training data at industrial scale.
Why this matters to Source
Source exists because much of the world's evidence remains physical, non digitised and inaccessible through ordinary search engines, which is why our vetted researchers physically enter archives, libraries, registries and private collections to retrieve primary sources that databases cannot reach.
That work continually reminds me that information and evidence are not synonymous.
Information can be copied, summarised, indexed and generated, whereas evidence has context, materiality, location and provenance, and those qualities frequently become important only when a researcher arrives with a question that nobody previously thought to ask.
There are archives that have not yet been catalogued, relationships that have not yet been reconstructed, inscriptions whose names remain unidentified, and future researchers who will approach familiar objects with methodologies that do not yet exist.
When we destroy a physical source, we therefore lose more than the information we failed to capture at the time. We also lose the possibility of asking that source a different question in the future. Our complete guide to archive research goes into why so much of this evidence still isn't digitised, and our piece on artwork provenance research looks at how the same loss plays out specifically for art.
This is ultimately what Get Off My Back is about, because the danger of destructive book scanning is not simply that paper is being discarded, but that we are becoming increasingly comfortable with a philosophy in which the informational value of an object is assumed to exhaust its cultural value.
Perhaps the defining image of book destruction in the twenty first century will therefore not be the spectacle of the bonfire, but something much quieter: a warehouse filled with books, a blade removing their spines, a scanner capturing page after page, and a system recording everything it has been designed to recognise while discarding the object that might one day have told us something it was never designed to see.
For centuries, we have worried about those who wanted to destroy books because they feared the knowledge contained within them.
The challenge before us now is more subtle, because we must ask what happens when we value the knowledge inside a book so highly that we convince ourselves the book itself has nothing left to say.
If there's a physical record you need checked, verified, or retrieved before it disappears into a scanner, you can post a request and get matched with a researcher who already has access to it.
