That doesn't mean they're crawling the URLs, just that they're indexing the content of PDFs/images for your search. Which is, honestly, a pretty useful feature. Whereas Facebook is doing this without providing any value to the user.
I've worked with email security appliances that will visit the pdf url when detonating the attachment in a Sandbox. Dynamic analysis helps find 0day's ,sometimes the link leads to something worse,the link many times is a url shortner too
