I wrote a paper this year (after about a 15 year hiatus) and apparently submitted papers, this was to an elsevier journal, now automatically go though a plagiarism detector of some kind that gives a score that can't exceed some value. We actually failed the first time because we'd published the paper on Arxiv first, which seems like a big flaw in the system. I don't have a lot of faith in, nor do I support automated detection like that, but it does happen systematically now.

Not without false positives, which are unacceptable (unless you're willing to put the work saved from plagiarism checking into absolving the innocent).

I was forced to submit essays to a plagarism checker in high school (well over a decade ago). The company running this checker had, in their infinite wisdom, included a check for writing complexity under the assumption that a 10th grader who writes like an adult is probably cheating. This was embarrassing for everyone.

As long as it's available to students so they can check their work for unintentional plagiarism.