The AI Rare Book Shredding Panic
Our read
The panic over AI companies shredding rare books is peak midwit theater. If a book is so rare that it only exists in three moldy basements, digitizing its contents into a neural network that can actually use it is a massive net positive for human knowledge, even if the paper gets recycled.
What happened
AI companies are reportedly sourcing rare, out-of-print, and physical books to scan and shred for training data, sparking outrage among archivists and collectors.
The brief
The outrage isn't about saving history; it is about hoarding it. Archivists would rather let a text rot in a locked vault than let an LLM make its contents useful to the public.
The sides
- The Preservationists
Physical books are sacred artifacts that must not be destroyed for machine training.
- The Accelerationists
Digitizing rare knowledge into a permanent model weights file is the ultimate form of preservation.
Why now
The physical destruction of rare books by AI labs has hit the timeline, triggering a massive wave of outrage from bibliophiles and tech skeptics who view the practice as a literal digital book burning.
Questions
Are AI companies actually shredding rare books to train models?
Yes, some AI labs buy physical out-of-print books, cut the spines off to run them through high-speed scanners, and recycle the paper. This destructive scanning is the fastest way to digitize rare text for training data, turning paper-bound knowledge into machine-readable tokens.
Why do AI labs destroy the books instead of just reading them?
Destructive scanning is a standard industry shortcut for high-volume digitization. Flatbed scanning of bound books is too slow and expensive, so labs slice the bindings to feed pages through automatic document feeders, trading the physical artifact for a perfect digital copy.
Is AI book shredding bad for the preservation of human history?
No, it actually preserves the information. Keeping a rare text locked in a moldy basement where nobody can read it is not preservation, it is hoarding. Digitizing the text into a neural network ensures the actual knowledge survives and remains useful long after the physical paper decays.
Who is actually buying and destroying these physical books for AI training?
Third-party data brokers and specialized scanning operations are the ones purchasing physical books in bulk to feed the AI pipeline. These middlemen buy cheap, out-of-print paperbacks and discarded library stock, slice the spines off for high-speed sheet-fed scanners, and sell the clean text files to AI labs. The tech giants themselves rarely touch the physical paper, outsourcing the manual labor to avoid the bad optics of running a literal industrial book shredder.
Why can't AI companies just use existing digital libraries like Google Books?
Copyright lawsuits and licensing restrictions have locked up the vast majority of existing digital archives. Google Books and the Internet Archive are constantly tied up in federal courts, making their databases a legal minefield for frontier model training. Sourcing physical copies of obscure, out-of-print books and scanning them directly allows AI companies to bypass digital rights management and build proprietary datasets with less immediate legal exposure.
What is the strongest argument against this high-speed scanning practice?
The real risk is the permanent loss of unique historical context that cannot be captured by a flat optical scan. Marginalia, physical binding techniques, paper watermarks, and even the chemical composition of ink hold vital historical data for researchers. When a rare or semi-rare volume is destroyed for its text, we lose the physical artifact forever, trading deep historical evidence for a flat string of tokens in a commercial database.
How does this scanning panic compare to historical preservation efforts?
This is simply the industrial-scale evolution of microfilming and early digital archiving, which have always involved destroying books to save their contents. Throughout the 20th century, libraries routinely destroyed newspapers and rare journals after filming them to save physical shelf space. The only difference now is the speed of the destruction and the fact that the end user is a commercial neural network rather than a university researcher.
What happens to the market value of physical books if this trend continues?
The market value of surviving physical books will likely skyrocket as physical paper becomes a premium status symbol. As more text is digitized and digested by AI, physical libraries will transition from utilitarian information repositories to high-end art galleries. Collectors will pay a massive premium for untouched, physical copies of books precisely because they represent a tangible, unhackable link to pre-AI human culture.
Receipts
Related dispatches
- The AI Midwit TrapThe corporate lie of AI as a harmless co-pilot is collapsing first for the people who actually use it. If your daily output can be replicated by a prompt engineered by a junior staffer, you are not being upskilled; you are being offboarded.
- The $1.5 Billion AI Absolution FeeSilicon Valley's grand strategy of 'move fast and break copyright' has officially entered the tax phase. Anthropic paying $1.5 billion isn't a defeat for AI; it is the cost of doing business to turn stolen intellectual property into clean, institutional-grade enterprise software.
- Elon Musk's AI OdysseyThe media's panic over Musk's 'AI remake' threat misses the real shift: the legacy entertainment pipeline has become so formulaic and hollow that a theoretical prompt-engineered epic actually sounds more compelling to audiences than another expensive, focus-grouped Hollywood disaster.
- The Fake Local Outrage EngineAdversaries are not inventing our divisions; they are just running a script on our pre-existing software. If your local energy policy is fragile enough to be derailed by a handful of AI-generated Facebook accounts complaining about noise, the problem is your system, not the bots.
- The Hollywood AI strike was a theater kid PR stuntThe public theater of the Hollywood strikes was never about stopping AI; it was about negotiating the price of the surrender. While actors and writers marched with cardboard signs for the cameras, the actual business of filmmaking quietly automated the tedious, expensive parts of the pipeline because math always wins over sentiment.
- The AI Silicon Squeeze is Killing Consumer GamingThe AI gold rush is cannibalizing consumer hardware, forcing console makers to kill physical media, delay next-gen specs, and monetize gameplay with programmatic ads just to survive the global memory shortage.