Get the Ricochet app
Ricochet will work fine in a browser on your mobile devices, but for the best experience please download our app.

3 Whisky Happy Hour
The Ricochet Podcast
GLoP Culture
Beyond the Polls with Henry Olsen
Gray Matters
Daily Signal
The Erick Erickson Show
Giving Ventures
Mom Wars
3 Martini Lunch
Chicks on The Right
Parsing Immigration Policy
Federalist Radio Hour
What the Hell Is Going On
Freedom to Learn
The Learning Curve
The Conservative Woman’s Guide
Heritage Explains
Not Dead Yet
Who Funds that?
The Radio Free Hillsdale Hour
The Glenn Show
Call Me Back - with Dan Senor
Hillsdale Dialogues
Law Talk with Epstein, Yoo & Cooke
The Libertarian
The Report Card with Nat Malkus
The Kevin Roberts Show with Larry O’Connor
The American Idea
Madison’s Notes
Conservative Crossroads with Henry Olsen
The Back of the Book
Martini Shot
The Dave Carter Show
The Powerline Show
In Dissent
Heard at Heritage
Smart Girl Politics
James Lileks’ The Diner
Case in Point: The Legal Show on the Hottest Legal Cases in Politics and Culture
JobMakers
One on One with Robert Doar
Homeschooling Journeys with Curious MikeRicochet will work fine in a browser on your mobile devices, but for the best experience please download our app.
Tired of the toxicity of social media? Why not join Ricochet for just $5/month? Think of it as an online bar. Come in, have a drink, hang out, say whatever you want. But don’t badger, bully, or berate the other drinkers. That's our only rule.
Book sellers across the country and the world are suddenly getting large orders for any kind of books, no matter the theme, published before 2022. The purchasers are hidden through a third party, but the sellers are talking to each other, and there is a pattern. They learned the books are supposedly being bought up in mass for AI Learning, ripped apart and scanned before being dumped. There are other concerns in this story.
AI companies are buying tons of old books because they're free of AI slop
ISBNdb, a company that sources printed books for AI companies to turn into training data, tells clients “the optics problem is real.” As AI companies search for more training data to improve their models, one company is offering old, printed books as an ideal source because they are guaranteed to be free of the very AI slop AI companies are producing. “The world's best AI training data is sitting on a shelf,” ISBNdb, a company that produces what it claims is “the world’s largest book database.”
I was able to read the above story on the stock button, but now it's behind a paywall. Here is another article from Fortune:
“I was shocked!” said de Vries, who provided Fortune with the email and subsequent spreadsheet containing 3,001 titles, mostly published between 2020 and 2021 by academic publishers like Emerald Publishing, Elsevier, Wiley, Routledge and Oxford University Press, and ranging from business and education to engineering, public policy and medicine. “Very unusual,” de Vries said. “That’s why I regarded the request as spam or phishing.”
De Vries wasn’t the only one. He told Fortune several other Dutch antiquarian booksellers received the same request and, like him, dismissed it. (They didn’t take “it seriously” and “considered it a scam,” he said). While the inquiry had little effect on his business, he said, it could have been a different story for booksellers who specialize in selling secondhand books at scale.
“I deal in old and rare books,” de Vries said, adding he rarely receives book orders by e-mail and if he does it’s at most for only one book. “These books are expensive and exclusive. So I do not do bulk trade.”
The Fortune story followed up on the original story from 404 (link above) and found denial, and the information linking the third party was removed:
Earlier this year, 404 Media reported that ISBNdb, a company best known for maintaining metadata tied to International Standard Book Numbers (ISBNs), the unique numerical identifiers encoded in the barcodes found on most books, had begun advertising services to help AI labs source bulk printed book purchases ranging from 1,000 to 1 million books. According to the report, the company’s website said the service was tailored to “LLM training needs” and “delivered at the scale AI demands.”
The webpages have since been removed. ISBNdb later told Fortune the service was never launched, it never purchased or scanned books for AI companies and the webpages reflected an exploratory concept rather than an active business. Similar reports have emerged in Germany and Switzerland, where local media reported that secondhand booksellers received unusual bulk orders for highly specialized books that appeared inconsistent with traditional collecting or resale. Zoom Books, a Canadian company identified in some of the reports, denied these allegations and told Swiss broadcaster SRF the purchases were part of its regular recycling and trading model.
Apparently from 2022 on, some book content has been generated by AI — establishing a counterproductive circuit leaving AI to train itself, and another claim is that there are hidden "disruptions or bugs" that disable AI from copying those books! What world are we now living in? I am beyond appalled by this. I love to read, I treasure both old and new books. I've never owned a Kindle - the real thing is the only way I read. Now there is a behind-the-scenes effort to limit what could be available to current and future generations and to eliminate the choice.
This story doesn't seem to be appearing in the mainstream. A judge named in the first article by 404 deemed this legal: that the books have outlived their use for the original writer to be paid — or some idiotic reasoning such as that.
This has a name. They are calling it Destructive Scanning. I call it insanity.
This post was promoted to the Main Feed at the recommendation of Ricochet's members.
Discussion (66)
To read this bio, please join Ricochet.
Modern book burning. Nice and quiet.
To read this bio, please join Ricochet.
One of the big AI companies just settled a lawsuit. They downloaded a million(?) pirated digital copies of books for training. They will end paying the authors $3,000 each. If the bought a physical copy and scanned it would have been cheaper, but the big AI companies do not care about cheap. They need quality training data fast. Obviously the Ricochet PTB should figure out how to monetize the content on this site.
CONTRIBUTOR
To read this bio, please join Ricochet.
There are a great many books that I could spare to the AI chompers. The Audacity of Hope, for example. Back in 2010 my library had five copies of that on the shelf and thirty holds on The Road to Serfdom. While I love books and I prefer physical copies to digital ones I do think we need to distinguish between chopping up some books and chopping up the last copy of a book. My worry here isn't so much that they're eating books as that they're going to scoop up underappreciated gems which will no longer be available for the diligent searcher.
There are already people who will sell you books by the foot, sorted by color or height, for the interior designer who likes the look of the things but can't be bothered with the contents. If those services got priced out of the market by AI pulpers I wouldn't cry very hard.
There are, as far as I know, no laws against using things for AI training data, and so the Silicon Valley types are doing as much as they can before the government catches up. We have a Congress for making laws. They should do something.
This problem is also made more acute by the length of our copyright laws. The more restrictive the regime the more likely it is that someone will buy up the last copy of it and destroy it without anyone legally being able to save the text.
CONTRIBUTOR
To read this bio, please join Ricochet.
The best explanation I have heard for these mysterious book buy is financial arbitrage. They are buying up books overpriced on Amazon, available at used bookstores which do not list at Amazon at lower prices for resale on Amazon.
Explanation here. The guy runs a used bookstore and has been selling books that fit this buying pattern.
CONTRIBUTOR
To read this bio, please join Ricochet.
Google had a project to digitize and make available the collected writings of mankind. As part of it they designed and patented fast, non-destructive book scanning machines. Hmm... a quick word on the mechanics.
Let's say you've got a book and you're looking to photocopy something out of it. You open the book to that page, you place it face down on the copier, you press the lid down as far as you can go, and you hit 'print'. You get a legible copy, but it curves up a bit towards the spine where you didn't press hard enough.
Now let's say you wanted to do that on pages by the million. At that scale you might look for mechanical ways to turn the pages, and digital ways to read the curved-up bits more accurately. Or you could slit the spine out, remove the covers, and run a stack of papers through, photographing both sides of each in turn. It's much simpler, much less work. That is, I understand, how the AI companies are doing it.
Google had a project to digitize knowledge, and they did so non-destructively. Then it was made radioactive by legal troubles, much like the ruins of a city after a nuclear war. But consider that. Google has already done the work to digitize a large enough sample of books to train AIs off of. If they could lease that legally to train AIs it'd be cheaper than destroying books by the cord. But legally speaking that information remains as radioactive as Ft. Knox after a visit from Auric Goldfinger.
You're seeing five of 66 comments. Sign in, or join for just $5 per month, to read them all and add your own.