There is many good reasons to "scan" _all_ content. How about spam, illegal content, or anything that violates their TOU? These days if you do not have a litany of algorithms performing these tasks, the worst case is that you'll end up like Facebook where genocide occured (in Myanmar) because the content was not in unicode and the detection systems couldn't "read" it.

Like any other site, accounts with less "trust" are probably going to be flagged by automatic mechanisms. I've been on a few teams with various companies that have done things like fraud detection and it almost always works this way. sometimes there is of course, legitimate false positives.