I am sure that in some cases it is completely intentional (be it a bot or a human) with an intent to be disruptive in some way, but in other cases I think it can be reasonably been seen as being genuine. Sometimes the respondent will eventually admit that they took it the wrong way. Often it comes from usernames who have enough of a reputation that you can have good reason to believe it is a real person with no specific motive.