[Quoth]

PDFs from scans, even if OCR added, ARE simply digital photocopies. No real library scans or photocopies any physical book still in copyright to loan it. If they do eBooks at all, they licence an appropriate quantity. It need not be one. Fair use involves only photocopying or scanning a portion. I expect the Google case to be reopened once it's obvious to rights holders and publishers that Google over time is sharing copies of the entirety of everything they scan and monetising it via search without a cent in royalties.

Most serious photocopiers are now actually computer scanners and laser printers in a box with a computer. Purely analogue photocopiers, even the dry toner kind, are candidates for museums.

You can't buy a DVD and offer digital loans without a per loan royalty and/or a contract with the rights holder. Same with ebooks.
It's not remotely the same as buying a CD, DVD, ebook and changing the format for your own gadget, even in countries where that is technically illegal due to DRM circumvention, no individual would be sued for personal use. But websites or operators "sharing" digital copies by any means and any conversion have always been successfully prosecuted or sued. It's only the level of damages that's ever in contention.