For a while, the Internet Archive built a book scanner that rests the book in a V-shaped cradle. A volunteer turns pages by hand and lowers/raises a pair of glass panes that gently press the pages for imaging by a pair of DSLR cameras on an angled mount. The whole assembly isn’t automatic, but can be easily operated by hand.
Interestingly the Internet Archive copied the open source design (of their scanner) from the site https://diybookscanner.org/ ( as they are allowed to by its open source licence ) The internet Archive then effectively refused to release any details back to the community. After a lot of “pushing”, the Internet Archive did acknowledge the source their design was based on came from the site bookscanner.org. This would have been very disappointing, in my opinion for the designer of the scanner – who released the info open source. At one time Internet Archive sold this scanner to organizations for $10K I think the price has dropped now ( I think ) to a few thousand.
I'm not sure if one of those machines was the cause, but I've seen far too many old books on archive.org which have pages that appear to have been torn by the scanner; thus I doubt they're manually doing it.
https://blog.archive.org/2021/02/09/meet-eliza-zhang-book-sc...