Fair enough, though another possibility I see is that the automated training process for LLMs used OCR for those papers (Or an already existing text version in the internet was using bad OCR) and those papers with the mashed word were written partially or fully by an LLM.

Either way, the blanket term “AI” sucks and it’s honestly getting kind of annoying. Same with how much LLMs are used.

Wouldn’t it be OCR in this case? At least the scanning?

Why the hostility? I get not liking AI (I do not like it either), but being hostile is not necessary. Especially since the Fediverse (Sorry if this is the wrong terminology) is meant to be open and allow communities of all kinds. If you do not like something, downvote it or just block the community it is from and move on, no need to be hostile about it.