A very spicy take I have is that the "wow suddenly there are fake citations in papers" phenomenon is less about AI and more like the way that trans women have lower breast cancer risk but a younger age of first discovery: once circumstances force you to check regularly you find things more often
Like, you get it, right? A culture of not checking what the fuck you were citing can't have arisen overnight. Maybe the phenomenon of incorrect names and titles showing up in bibliographies is new, but it definitely means that there were A Lot of "fake citations"-qua-"unread papers unrelated to the claim they supposedly support" already.

