In an Age of AI, Who Will Protect Our Human Stories?
Recent reports about AI companies buying physical books, cutting off their bindings and scanning the pages have touched a nerve for a simple reason: books are never just containers of text.
The basic practice is not new. Libraries, archives, universities and data companies have digitised books for decades. What feels different now is the scale, speed and purpose. Some companies involved in artificial intelligence need very large collections of written material to train models that can read, summarise, translate, answer questions and generate new text. One way to build those datasets is to scan printed books.
Many of the books involved appear to be common titles, bought legally through ordinary channels such as second-hand sellers, wholesalers or remainder markets. In many cases, destructive scanning is used because it is faster and cheaper. The binding is removed, the pages are fed through scanners, and the book becomes a digital file rather than a physical object.
That fact alone does not make the practice sinister. It does raise a question that deserves more care than outrage: what do we lose when a book is treated only as information?
Some rare book dealers, librarians and preservation experts have expressed concern that uncommon, out-of-print or harder-to-find books may be destroyed in the same process. That concern is not the same as proof of widespread loss of rare cultural material. It is a warning from people who understand how fragile the world of books can be. A title may look ordinary until someone tries to replace it. A modest local history, a small-press poetry collection, an early edition, a marked-up family Bible or a cookbook from a regional community group can carry value that a scanner cannot fully capture.
The issue, then, is not whether AI is good or bad. Artificial intelligence is already helping people learn languages, analyse medical data, improve accessibility, support research and work across barriers of time, distance and disability. The more useful question is how society can welcome powerful new tools while still protecting the human inheritance that made those tools possible.
Recent scanning reports raise a practical preservation question
Destructive scanning has a straightforward appeal. A bound book is awkward for a machine. Pages curve near the spine. Text can vanish into the gutter. Turning pages by hand takes time. Removing the binding lets scanners capture pages quickly and evenly.
For mass digitisation, that efficiency matters. A company scanning thousands or millions of pages will naturally seek speed, consistency and low cost. Common books, especially damaged or excess copies, may seem like reasonable candidates. If a paperback novel or textbook exists in large numbers, destroying one copy to create a digital version may not feel like a cultural loss.
The concern begins where certainty ends.
Book markets are uneven. Some books that appear ordinary are not easy to find. Small print runs, regional publishing, old technical manuals, independent literary work, religious texts, school readers, local histories and specialist books can slip through the cracks. They may not be rare in the grand museum sense, but they can still be important to one community, one field of study, one family or one future researcher.
Preservation experts often think in terms of risk. A single destroyed copy may not matter. A pattern of casual destruction might. The difference depends on what is being scanned, how many copies remain, whether libraries hold preservation copies, and whether the physical features of the book have value beyond the printed words.
That last point is easy to overlook. A scanned text can capture sentences. It may capture page layout, illustrations and marginal notes if the process is careful. It does not always capture binding, paper quality, repairs, ownership marks, inscriptions, stains, pressed flowers, bookmarks or the way an object has travelled through human hands.
A digital copy can be useful. A physical book can be evidence.
Books contain more than words
A book begins as thought, but it becomes more than thought by the time it reaches a reader. It carries labour from writers, editors, designers, printers, booksellers, librarians and readers. It holds choices about language, pacing, paper, type, cover art and structure. It reflects the culture that produced it and the private worlds of the people who kept it.
This is easy to see in literature. A novel may contain years of observation, grief, humour, patience and invention. A memoir may hold a life that would otherwise disappear. A poetry collection may preserve ways of feeling that are difficult to explain in ordinary speech.
Yet the same is true of less celebrated books.
An old cookbook is rarely just a list of recipes. It might show how families stretched ingredients during lean years, how migrants adapted dishes to local shops, how a grandmother changed measurements by instinct, or how a community raised money through shared food. A Bible or prayer book may hold births, deaths, marriages and pencilled notes from people who turned to it in sorrow or gratitude. A field guide may show the marks of a child learning to name birds, shells or wildflowers. A journal may hold the private weather of a person’s mind across decades.
Even a schoolbook can become a time capsule. A name written inside the cover. A doodle in the margin. A pressed leaf between pages. A sentence underlined by someone who was young when they read it, and old when they found it again.
Digital text can preserve the words. It may even preserve an image of the page. But the object still carries a different kind of meaning. It reminds us that knowledge did not appear from nowhere. It passed through bodies, places, homes, shelves, bags, classrooms and hands.
Books are records of attention. They show what people cared enough to write, print, buy, borrow, annotate, lend, hide, repair and keep.
The sentimental value of physical books still matters
Sentiment is sometimes treated as a weakness in discussions about technology. It should not be. Sentiment is one way humans assign value to things that cannot be measured easily.
A battered children’s book may have no resale value and still be priceless. It may carry the memory of a parent’s voice at bedtime, a grandparent’s lap, a classroom corner or a hospital room. A family cookbook with splash marks and handwritten changes can mean more than a perfect digital recipe. A collection of paperbacks gathered over a lifetime can show the shape of a person’s curiosity.
In Australia, as elsewhere, many homes hold books that are part of family history. They may include migration stories, church or temple texts, local histories, gardening books suited to a particular climate, war memoirs, poetry anthologies, community cookbooks, technical manuals, children’s classics and photo albums stored on the same shelves. Some families keep books in languages their younger generations do not fully read, but the books still matter. They are proof of origin, movement and belonging.
The sentimental value of books does not mean every copy must be preserved forever. Homes have limits. Libraries have budgets. Archives must make hard choices. Still, sentiment reminds us that value is not always visible to the market.
A rare book dealer may see scarcity. A librarian may see research value. A historian may see evidence. A family may see love.
All of these forms of value can be real at once.
Physical books offer quiet comfort in a digital world
Digital reading has made knowledge more accessible than at any point in human history. A person can carry a library on a small device. Search functions help students and researchers find passages in seconds. Audiobooks support people who cannot easily read print. Ebooks can enlarge text, change contrast and travel across the world instantly.
These are genuine gains. They should not be dismissed.
Still, many people continue to find comfort in physical books. The comfort is quiet, almost ordinary. The weight of a book in the hand. The soft resistance of a page turning. The smell of paper in a second-hand shop. The sight of a favourite spine on a shelf. The memory of where a passage sits on the page. The small ritual of choosing a bookmark.
These experiences may seem minor, but they shape how people read. A printed book asks for a certain kind of attention. It does not flash with notifications. It does not update while we hold it. It keeps its place. It gives the mind a stable object in a world of constant motion.
For some readers, that stability is part of the pleasure. For others, it is part of the relief. A physical book can slow a room down. It can make private thought feel possible again.
Preserving physical books, then, is not about rejecting digital life. It is about keeping tangible connections to our shared human story. A scanned page can travel further than a printed one. A printed book can remind us that knowledge once had a texture, a smell, a weight and a place on someone’s shelf.
Both forms can serve human needs. Neither has to erase the other.
Digital preservation is valuable and AI can serve society
The case for digitisation is strong. Digital copies can protect access when physical objects are fragile, distant or at risk. They can help scholars compare texts. They can let people in regional areas read materials held in major city libraries. They can preserve books threatened by flood, fire, mould, neglect or political instability.
Digital access can also broaden cultural memory. A local history locked in one reading room serves fewer people than a careful scan that can be searched and shared. A fragile newspaper, once digitised, can become available to family historians and researchers across the country. A public-domain novel can reach new readers who would never encounter it in print.
AI adds another layer of possibility. Models trained on human knowledge can help translate texts, summarise archives, describe images for blind and low-vision readers, support literacy, assist researchers and help people navigate complex information. Used well, these tools can extend the reach of human learning.
They can also reveal patterns too large for any one person to see. A researcher might use AI to search across old newspapers for references to a place, a disease, a phrase or a forgotten public figure. A student might use it to understand difficult historical language. A person with limited mobility might use it to explore collections that would otherwise remain out of reach.
These benefits matter. They are part of why the conversation must stay balanced. Treating AI as an enemy would miss the real promise of the technology. Treating books as raw material would miss the real depth of culture.
The better path is careful choice.
Common duplicate books may be reasonable candidates for destructive scanning, especially if they are damaged or widely held. Uncommon or culturally valuable works deserve more caution. Non-destructive scanning, professional conservation, library partnerships and archival review can help decide which books can be digitised quickly and which need protection as physical objects.
This kind of care may cost more. Yet cultural preservation has always required patience. A society that can build powerful AI systems can also build better habits around the sources those systems learn from.
Some books deserve protection beyond their text
Not all books need the same treatment. A mass-market paperback with millions of copies in circulation is different from a signed first edition, a small-run community history or a nineteenth-century school primer with student notes. A damaged duplicate held by many libraries is different from a scarce book with uncertain holdings.
A practical preservation approach might ask several questions before a book is destroyed for scanning:
How many copies are known to exist in libraries or archives?
Is the book out of print or difficult to replace?
Does it have local, cultural, religious or historical significance?
Does the physical object contain inscriptions, annotations or ownership marks?
Could non-destructive scanning capture it safely?
Would a library, archive or specialist bookseller recognise value that a bulk buyer might miss?
These questions do not require panic. They require judgement.
Libraries already make similar decisions. Archivists decide what to keep, what to digitise, what to conserve and what to let go. Rare book dealers identify editions, bindings, provenance and scarcity. Community historians recognise material that larger institutions may overlook. Indigenous communities, migrant communities and religious communities may understand the significance of texts that outsiders treat as ordinary.
A respectful system would bring these forms of knowledge together. It would not assume that market availability equals cultural disposability.
There is another reason to protect some physical books: future technologies may learn from things current scanning misses. Better imaging may one day reveal erased marks, paper composition, printing methods, repairs or hidden ownership evidence. Destroying the object removes that future possibility.
A scan can be an opening. Destruction is final.
AI-generated content may make human creativity more valuable
The discussion does not end with scanning books. Artificial intelligence is now able to produce fluent text, polished images, music, code and video. Some of this work is useful. Some is beautiful in its own way. Some is bland. Some is misleading. Much of it will become ordinary.
As machine-made content grows, authentic human creativity may become more valuable, not less. People may seek work that carries lived experience, moral struggle, humour, risk, memory and embodied knowledge. They may want to know who made something, why it was made and what life shaped it.
A poem written after loss carries a different meaning from a poem assembled from patterns in data. A painting made by someone working through illness, joy or exile holds a human story beyond the image. A song sung in a kitchen by an elder may matter even if it is never professionally recorded. A handwritten letter can feel alive because the hand moved across the page.
AI can imitate patterns of expression. It can remix styles and generate convincing forms. It does not grow up in a household. It does not miss a parent. It does not wait at a hospital bed, migrate across oceans, raise a child, lose faith, regain hope, plant a garden or forgive an old friend.
That does not make AI useless. It makes human experience distinct.
The rise of generated content should encourage more people to create, not fewer. Write the story. Keep the journal. Record the family history. Make the painting. Learn the instrument. Read aloud to a child. Ask an elder about their childhood. Save the recipes that were never written down. Name the people in old photographs. Keep letters. Visit libraries. Buy books when possible. Borrow them when buying is not possible. Talk to people who remember what no database holds.
Human creativity is not only the production of content. It is the act of paying attention to life and shaping that attention into something another person can receive.
Wisdom is not the same as information
AI systems can process extraordinary amounts of information. They can identify patterns, summarise material and answer questions at remarkable speed. This can make them useful companions in learning. Yet wisdom is different from access.
Information can tell us what happened. Wisdom asks what it meant, who was affected, what should be remembered and how we ought to act.
Wisdom grows through lived experience, empathy, curiosity and relationships. It comes from making mistakes and taking responsibility. It comes from listening to people whose lives differ from ours. It comes from grief, love, work, care, conflict, apology and time. It comes from reading, too, but not only from reading. The book matters because it meets a person in the middle of a life.
This distinction matters for the future of AI. If society treats knowledge as a warehouse of extractable data, it may gain speed while losing depth. If society treats knowledge as a human inheritance, it can build tools that serve learning without flattening its sources.
A model can answer a question about a war. A veteran’s diary can show fear, boredom, confusion and tenderness. A model can summarise a migration policy. A family story can reveal what it felt like to leave one home and build another. A model can describe a recipe. A grandmother can show the pinch, the smell, the timing and the reason the dish mattered.
The future needs both technical intelligence and human wisdom. One cannot replace the other.
A better future protects both innovation and inheritance
The choice ahead is not between AI and books. It is between careless progress and careful progress.
Careless progress treats every available object as input. Careful progress asks what should be scanned, how it should be scanned, who should be consulted and what must remain. Careless progress assumes that if words have been captured, the work is done. Careful progress understands that culture lives in objects, contexts, communities and memories.
A hopeful path is possible.
AI companies can work with libraries, archives, publishers, authors and preservation experts to build better practices. They can avoid destroying uncommon books when non-destructive methods are available. They can support digitisation projects that preserve access without erasing physical heritage. They can document what they scan and how they handle materials. They can recognise that responsible data gathering is part of responsible technology.
Libraries and archives can keep expanding public access while protecting physical collections. Independent bookshops and second-hand dealers can continue rescuing books from obscurity. Readers can support authors, borrow from libraries, donate thoughtfully, preserve family books and pay attention to the stories in their own homes.
None of this requires nostalgia to defeat technology. It asks technology to remember its debt to human culture.
The reports about destructive book scanning matter because they point to a larger truth. AI learns from human expression. That expression came from lives, communities, languages, histories and imaginations. The tools may be new, but the source is ancient and deeply human.
AI and Books Why Human Creativity Still Matters is not just a title. It is a reminder that progress should widen our respect for human creation, not narrow it.
The future can hold powerful artificial intelligence and beloved physical books. It can hold searchable archives and handwritten notes. It can hold digital access and quiet shelves. It can hold machine assistance and original human thought.
What matters is the care we bring to the transition.
Support authors. Use libraries. Visit independent bookshops. Value archives. Save the family stories before they fade. Keep making things that come from lived experience. Read old books and new ones. Ask better questions of technology, not because technology is the enemy, but because human creativity is worth protecting.
If future generations inherit only our information, they will inherit less than we were. If they inherit our knowledge with the humanity behind it, they will receive something far richer: the record of a species that kept learning, imagining, remembering and reaching for one another.
Comments