Books like that are pretty much all AI trainers are interested in.

The classics are all in public domain already so there's no reason for them to want to destroy a rare edition of Charles Dickens or Arthur Conan Doyle. And the stuff that isn't in public domain went through an era of mass publishing where copies are a plenty.

But further, just the nature of LLM trainers is they probably aren't particularly interested in lesser known books from the 30s or 40s. Why would they want, for example, an Chemistry book from the 30s? Something published in 2026 is much more up to date and likely to be useful. The only books they'd be interested in that period would be popular and have cultural impact. Meaning, there's probably plenty of copies laying around.