As AI tech companies increasingly buy and destroy books to feed to their AI models, Anna’s Archive is calling for volunteers to help preserve them for the public record.
As AI tech companies increasingly buy and destroy books to feed to their AI models, Anna’s Archive is calling for volunteers to help preserve them for the public record.
There is a glaringly obvious issue with this article that I see a lot of people in the comments seem to have missed:
Only one copy of any given book exists The pretense to this entire thing is that AI companies are destroying the books both because of the process used AND to prevent competitors from accessing the book. There is more than ONE copy of that book. If you buy a book, that doesn’t prevent me from also buying that book. You can not like AI for a number of reasons, but this is just stupid.
NOTE: If destroying the book actually prevented competitors from accessing it, that means that all of these AI companies are only going out on the weekends to used bookstores looking for good deals on second hand books.
There are plenty of books out there that don’t have many copies.
If you can purchase a book at a bookstore, then it can be assumed there are sufficient copies in production. They aren’t buying rare books here.
That’s exactly what I’ve heard they’re doing.
So there have been a bunch of articles about this lately. They play loose with descriptions in their titles. Here is one example: https://futurism.com/artificial-intelligence/ai-companies-destroying-rare-books
“Antique books”. One, notice they don’t mention any actual title or ISBN? Two, these are bought through a book reseller (like Book Depot or ThriftBooks), meaning that “antique” just means “used”.
In my personal library collection I have new books, I have used books, and I have rare books. These are not the same thing.
That seems like a leap that requires more evidence.