AI Training and the Destruction of Rare Books: Is Technology Creating a New Information Monopoly?

Introduction

Artificial intelligence is often presented as a technology that can preserve knowledge, make education accessible and help humanity solve difficult problems. But what happens when the race to build better AI starts consuming the very books, literature and cultural material that carry that knowledge?

A recent investigation by 404 Media has raised exactly this uncomfortable question. Reporters placed an Apple AirTag inside a rare book that was part of a large order suspected to be connected to AI training. They then followed the book's journey across the United States until it reached an Amazon facility in Las Vegas.

The discovery is remarkable because the book apparently ended up at a facility where workers reportedly scan and cut books. The cutting process involves removing the spine so that the pages can be processed more efficiently.

The story is not simply about one book or one Amazon warehouse. It points towards a much bigger transformation: the race among AI companies to acquire as much information as possible, convert it into digital data and gain an advantage over competitors.

And that raises a difficult question: when knowledge becomes valuable training data, who ultimately owns it?

1. How 404 Media Tracked a Rare Book

The investigation began after booksellers noticed something unusual in the rare-book market.

According to the transcript, sellers were experiencing an extraordinary increase in demand. AI companies were suspected of purchasing books in large quantities, sometimes without appearing particularly concerned about price.

One bookseller reportedly received an order for around 1,000 books through Biblio, a marketplace used for books.

404 Media saw an opportunity to find out where these books were actually going.

The seller agreed to place an Apple AirTag inside one of the books.

That small tracking device effectively turned an ordinary rare book into a travelling witness.

2. The Book's Journey Across America

The tracked book first travelled to California.

It then went by plane to Wisconsin before reaching a shipping and distribution facility near Kenosha Regional Airport.

The book remained there for roughly two weeks.

Eventually, it was picked up by what appeared to be a truck and began travelling across the American highway system.

Its journey continued through Colorado, including an overnight stop at a trucking travel centre.

Eventually, the AirTag showed that the book had reached its destination:

Amazon's LAS8 facility in Las Vegas.

This was the crucial discovery that connected the rare-book purchase to an Amazon operation associated with book processing.

3. What Happens Inside the Amazon Facility?

Online discussions among Amazon workers provided another important piece of the puzzle.

According to the investigation described in the transcript, the northern side of the LAS8 location housed another Amazon operation identified as VGT3.

One worker reportedly described the operation simply as scanning books.

Workers allegedly had different assignments. Some received books and scanned their barcodes, while others were assigned to cut books.

That leads to an obvious question.

Why Would Anyone Cut a Book?

If the goal is simply to read the contents, destroying the physical book sounds completely unnecessary.

But for large-scale digitisation, the situation is different.

A book can be processed page by page, converted into digital information and added to a database.

Cutting the spine can make the scanning process faster and cheaper because individual pages can be fed through scanning equipment efficiently.

In other words, something that looks like destruction from a librarian's perspective can look like an efficiency improvement from a data-processing perspective.

And that difference in perspective is at the heart of this controversy.

4. Why Are AI Companies Interested in Books?

Modern AI systems require enormous quantities of information.

Books contain something that is difficult to replace: carefully organised human knowledge.

A book may contain history, scientific explanations, philosophy, fiction, cultural observations, technical information or specialised knowledge that may not exist elsewhere in exactly the same form.

For an AI company, millions of books represent an enormous potential source of training data.

The more information an AI system can process, the more capable its developers hope it will become.

This has created a new form of competition.

Information Has Become a Strategic Asset

For decades, companies competed primarily through factories, machinery, patents, distribution networks and financial capital.

Today, data is another major competitive advantage.

AI companies are competing not only to build better models but also to obtain better and larger datasets.

That makes previously overlooked information suddenly valuable.

A rare book sitting quietly on a shelf may have little commercial value to most people.

But if its contents are unavailable digitally, the book could represent a unique information source.

5. The Strange Economics of Destroying a Rare Book

There is an almost paradoxical aspect to this story.

Imagine owning one of only a few surviving copies of a book.

Normally, its rarity makes the physical object more valuable.

But after someone scans it and destroys the copy, the physical book becomes even rarer.

The irony is striking: the information survives digitally while the physical object disappears.

This creates a strange economic equation.

The AI company gets the information.

The world loses another physical copy.

And whoever controls the digital database may eventually control access to the information in a way that the original book owner never could.

6. The Anthropic Copyright Case Adds Another Dimension

The transcript also refers to a lawsuit involving authors and Anthropic.

A judge ruled that Anthropic's scanning of books for training purposes could qualify as fair use under the circumstances described in the case, with the destruction of the original copies being one factor in the legal reasoning.

This is important because it demonstrates how traditional copyright concepts are being tested by AI technology.

Copyright law developed in a world where copying a book generally meant producing another copy.

AI changes the equation.

A company can acquire a physical book, digitise its contents, process the information through an AI system and potentially use the resulting capabilities at enormous scale.

The legal question is therefore no longer simply:

"Did someone copy a book?"

It becomes:

"What happened to that copy, why was it made, and what commercial use was ultimately made of the information?"

7. Amazon's Response

Amazon reportedly told 404 Media that it purchases books through commercial channels to help develop and improve products and services used by customers.

However, according to the transcript, the company did not answer several specific questions concerning the cutting of books, the number of similar facilities around the world and the public backlash over the destruction of books.

That leaves important questions unanswered.

How many books are processed this way?

How many facilities perform similar work?

Are rare books being treated differently from ordinary books?

And most importantly, how much of this activity is directly related to AI training?

These questions matter because scale changes everything.

Destroying one unwanted book is one thing.

Systematically acquiring and dismantling rare books for data extraction would be something much larger.

8. This Is Bigger Than Amazon

It would be easy to reduce the entire controversy to Amazon.

That would miss the larger issue.

The transcript describes a broader transformation in which technology companies are trying to capture information from books, artwork, films, television programmes and other cultural works.

The central issue is control over information.

Physical media once gave creators and consumers a certain degree of independence.

A DVD existed in your home.

A printed book sat on your shelf.

A purchased album remained in your collection.

Digital platforms changed this relationship.

9. The Problem With a World of Streaming

Consider a film that you bought on a streaming service.

One day, the company decides to remove it.

You cannot watch it anymore.

Perhaps another platform eventually gets the rights. Perhaps nobody does.

If you never owned a physical copy, your access depended entirely on someone else's platform.

This is one of the hidden consequences of the move from physical media to digital services.

Platforms can determine what remains available.

They can promote certain works through algorithms.

They can bury others.

And they can remove content altogether.

Physical media, despite its inconveniences, provides a form of permanence.

10. Books Are Different Because They Preserve Cultural Memory

Books occupy an especially important position because they are not merely entertainment products.

They preserve civilisation's memory.

A book can contain an argument from a scholar who died decades ago. It can document a culture, preserve a language or record an experience that may otherwise disappear.

When a rare book disappears, the loss is not necessarily financial.

It can also be cultural.

Digitisation Can Preserve Knowledge — But Who Controls It?

There is nothing inherently wrong with digitising books.

In fact, digitisation can be enormously beneficial.

A fragile historical book can be scanned and made available to researchers around the world without repeatedly handling the original.

The concern arises when digitisation becomes part of a commercial race in which the physical source material is destroyed and the resulting digital information becomes proprietary.

Preservation and extraction are not necessarily the same thing.

11. The Information Monopoly Problem

This leads to perhaps the most important question raised by the story.

What happens when a small number of powerful technology companies obtain enormous collections of information that ordinary people cannot access?

Imagine a library where the books are gone but a corporation owns the database containing their contents.

The corporation then builds an AI system capable of answering questions based on that information.

Users may not need to read the original books.

They simply ask the AI.

That sounds convenient.

But convenience can come with a price.

The AI company becomes a gatekeeper between people and knowledge.

12. AI and the Future of Education

The transcript connects the book controversy with another major development: the increasing integration of AI into universities.

The argument is straightforward.

If AI can explain history, literature, science and philosophy instantly, why maintain traditional educational structures?

But that question can become dangerous when efficiency replaces education itself.

Education is not simply the transfer of information.

A university teaches students how to question information, compare arguments, examine evidence and develop independent judgement.

An AI system can provide answers.

That does not automatically mean it can replace the intellectual process of learning.

13. The Texas Tech Controversy

The transcript also discusses controversy surrounding changes at Texas Tech University.

According to the New York Times reporting referenced in the discussion, university leadership was using AI-related systems to examine books and other educational materials for topics involving sexual orientation and gender identity.

Professors reportedly described the process as deeply troubling.

One example involved a course dealing with the Holocaust and literature.

The professor was concerned that references to gay and bisexual men being targeted by the Nazis could create problems under the university's new restrictions.

This illustrates a larger danger.

Technology does not automatically make decision-making neutral.

AI Can Reflect the Bias of Its Instructions

An AI system analyses whatever criteria humans give it.

If the people designing the criteria are politically or ideologically motivated, the technology can become an incredibly efficient tool for enforcing those preferences.

Instead of one administrator reading thousands of books, an AI system could potentially flag them at scale.

That is much more powerful.

14. What Happens to Academic Freedom?

Another example in the transcript involves teaching historical figures and subjects related to gender.

A medieval historian reportedly became concerned about discussing Joan of Arc because of historical references to cross-dressing.

Another professor reportedly removed material involving gender-changing figures from Viking mythology because of uncertainty over the university's directives.

This creates what could be called a chilling effect.

Professors do not need to be explicitly told, "Do not teach this."

If the rules are vague and the consequences appear serious, they may simply avoid controversial material.

The result can be self-censorship.

15. The Danger of Letting Technology Become the Gatekeeper

The most disturbing aspect of these developments is not necessarily the technology itself.

It is who controls it.

AI can help preserve knowledge.

AI can help students learn.

AI can make information easier to search.

AI can digitise fragile historical material.

But the same tools can also classify, filter, restrict and prioritise information.

The difference lies in the governance.

Who decides what gets included?

Who decides what gets removed?

Who owns the underlying data?

Who controls the algorithm?

And can ordinary people independently verify what the AI tells them?

16. Is Destroying Books the Modern Version of Book Burning?

The comparison with book burning is deliberately provocative, but it captures something important.

Traditional book burning destroys information by eliminating the physical source.

The modern scenario is different.

The information may be copied before the physical object disappears.

So technically, the knowledge has not been destroyed.

But something else has happened: ownership and access may have shifted.

Instead of destroying the information, it can be absorbed into a privately controlled technological system.

That is why the metaphor is powerful even though the two situations are not identical.

17. The Real Battle Is Over Control of Knowledge

The rare-book story ultimately points towards a much bigger competition.

AI companies need information.

Publishers want to protect intellectual property.

Authors want compensation and recognition.

Universities want to modernise while preserving academic independence.

Technology companies want massive datasets.

Consumers want cheap and convenient access.

And society wants knowledge to remain available.

These interests do not always align.

The coming years will therefore involve difficult debates about copyright, ownership, AI training, digitisation, preservation and access.

18. Why Physical Books Still Matter

In an increasingly digital world, physical books can seem outdated.

But their value may actually increase.

A physical book cannot suddenly disappear from your digital library because a platform changes its licensing agreement.

It does not require a subscription.

It does not depend on an algorithm.

It does not need a company's permission to remain on your shelf.

For rare and historically significant books, physical preservation becomes even more important.

Digital copies are valuable.

But they should ideally complement physical preservation rather than replace it through destruction.

19. What Should Happen Next?

The solution is not to stop AI development.

That would be unrealistic and potentially counterproductive.

Instead, society needs clearer rules around the acquisition and processing of cultural material.

AI companies should face transparent requirements concerning what datasets they use.

Researchers and the public should have greater visibility into how copyrighted and rare works are obtained.

Libraries, archives and cultural institutions should receive stronger protection.

And when historically significant physical works are digitised, preservation should be considered alongside commercial efficiency.

Transparency Must Come Before Trust

Technology companies frequently ask the public to trust them.

But trust works better when information is available.

If a company is purchasing thousands of books, people should be able to understand why.

If books are being cut apart, the process deserves explanation.

If AI systems are analysing educational materials, educators should understand the criteria being used.

Transparency is not an obstacle to innovation.

It is what makes innovation socially sustainable.

20. Conclusion

The story of the AirTag-equipped rare book is fascinating because it turns an abstract debate about AI into something tangible.

A physical book travelled thousands of miles, moved through warehouses and eventually reached an Amazon facility where books were reportedly scanned and, in some cases, cut apart.

That single journey raises much bigger questions about the future of knowledge.

AI companies are in a race for information. Books, art, films, academic material and other forms of human creativity have suddenly become valuable raw material for machine intelligence.

The danger is not simply that books might be destroyed.

The deeper concern is that knowledge itself could become increasingly concentrated in the hands of a small number of technology companies.

A society that moves from libraries and bookshelves to AI interfaces must ask an important question: are we merely making knowledge easier to access, or are we handing control of knowledge to new gatekeepers?

AI can become one of humanity's greatest tools for preserving and spreading knowledge.

But if the pursuit of technological advantage turns cultural memory into proprietary data, the price of convenience could be far higher than we realise.

The books may still exist inside the machines.

The question is whether we will still control the knowledge they contain.

Frequently Asked Questions

1. Why are AI companies interested in rare books?

Rare books can contain information that is difficult to find elsewhere in digital form. AI companies may value such material as potential training data because large and diverse datasets can help develop AI systems.

2. Why would an AI company cut a book?

Cutting the spine can make individual pages easier and faster to process with scanning equipment. It can therefore reduce the time and cost involved in digitising large numbers of books.

3. Did 404 Media prove that Amazon is destroying books specifically for AI training?

The investigation tracked a rare book from a large order to an Amazon facility associated with book scanning and processing. Amazon said it purchases books through commercial channels to help develop and improve products and services. The precise purpose and scale of the book-cutting operation remain important questions.

4. Is digitising books necessarily bad?

No. Digitisation can preserve fragile books and make knowledge accessible to researchers and readers worldwide. The controversy arises when physical works are destroyed, ownership becomes concentrated and the resulting digital information is used commercially without sufficient transparency or compensation.

5. Could AI eventually replace universities and teachers?

AI can automate some educational tasks and provide useful explanations, but education involves more than retrieving information. Critical thinking, discussion, research, mentorship and independent judgement remain important parts of learning.

Meta Description:
AI companies are racing to acquire books and information for AI training. Explore the rare-book tracking investigation, book destruction controversy, copyright concerns and the growing battle for control of knowledge.