I’ve been cataloging various AI sins in this space for a while, but few get my goat as viscerally as the news that Anthropic destroys books in the process of scanning them.
Anthropic was at the center of another controversy this week after details emerged from a recently settled lawsuit, revealing that the AI company had purchased rare books in bulk, scanned them using high-speed machines that sliced off their spines, and then shredded the original copies.
According to the settled lawsuit, Anthropic used a hydraulic powered cutting machine to neatly remove the pages from the books it procured from book resellers and then scanned them using industrial-grade imaging equipment, tech news site Futurism reported.
Evidently this destructive project went under the name Project Panama.
Vendor proposals cited in filings describe plans to process between 500,000 and 2 million books over six months, highlighting the extraordinary scale of the operation — on the order of tens of millions of dollars in book purchases, logistics, and scanning services.
Reportedly some of the books Athropic was buying and shredding were very rare indeed. It’s one thing to destroy common paperback reprints as part of the scanning process (no one will morn the loss of a paperback of Linda Goodman’s Sun Signs 1974), quite another to destroy rare and unique works as part of a project of cultural vandalism.
Shades of Vernor Vinge’s Rainbows End, where a tech company shreds books in the process of scanning them. If Anthropic took inspiration from that, they must have missed the part where the entire destruction process was unnecessary, as non-destructive methods were just as effective. Just as in the real world, there are a wide variety of non-destrictive overhead scanners. With an estimated post-IPO value of $500 billion to $1 trillion, I think Anthropic can afford to hire a few more people to run the scanners.
As a book collector and bookseller, this really angries up my blood.
Books are the primary vessels by which civilization is carried into the future. Intentionally destroying them to feed your pet Frankenstein is an act of inter-generational vandalism that slowly impoverishes the human race.
Books offer far superior memory retention to any other knowledge storage system thus devised. Go back to the early days of this blog a mere 16 years ago, and you’ll find linkrot has claimed most of the links I was talking about in 2009. Walk into most people’s houses these days, and I’d bet not many have working cassette players hooked up any more, much less something to play Edison’s wax cylinders. But a First Folio of Hamlet from 1623 can still be read today, once you can wrap your mind around the funky “f” being an s. As Bruce Sterling observed all the way back in 1991:
There’s talk nowadays in publishing circles about a new device for books, called a ReadMan. Like a Walkman only you carry it in your hands like this…. Has a very nice little graphics screen, theoretically, a high-definition thing, very legible…. And you play your books on it…. You buy the book as a floppy and you stick it in… And just think, wow you can even have graphics with your book… you can have music, you can have a soundtrack…. Narration…. Animated illustrations… Multimedia… it can even be interactive…. It’s the New Hollywood for Publisher’s Row, and at last books can aspire to the exalted condition of movies and cartoons and TV and computer games…. And just think when the ReadMan goes obsolete, all the product that was written for it will be blessedly gone forever!!! Erased from the memory of mankind!
Now I’m the farthest thing from a Luddite ladies and gentlemen, but when I contemplate this particular technical marvel my author’s blood runs cold… It’s really hard for books to compete with other multisensory media, with modern electronic media, and this is supposed to be the panacea for withering literature, but from the marrow of my bones I say get that fucking little sarcophagus away from me. For God’s sake don’t put my books into the Thomas Edison kinetoscope. Don’t put me into the stereograph, don’t write me on the wax cylinder, don’t tie my words and my thoughts to the fate of a piece of hardware, because hardware is even more mortal than I am, and I’m a hell of a lot more mortal than I care to be. Mortality is one good reason why I’m writing books in the first place.
Earlier this year Sony announced that it was deleting 550 movies from the libraries of people that bought them. But nobody’s deleting books from my library. Permanence is one area where analog books are far superior to their digital counterparts.
Books are enduring artifacts of human achievement and means of transmitting knowledge across generations. I own books from the libraries of famous writers and collectors, from a Henry Kuttner novel inscribed to Robert Bloch to an inscribed association copy of H. G. Wells’ The Food of the Gods. Some day after I’m gone, these books will pass into the libraries of other readers and collectors, each a link in the long chain of cultural memory.
A company that wantonly destroys books as part of its schemes of AI domination can’t, and shouldn’t, be trusted to do anything else with integrity.