Well if they make AI to detect and remove bots there will be bots with AI to avoid detection, and naturally there is always the possibility of false positives, there is always an arms race. I hate spam, there are figures that estimate 80% of email traffic on the web is spam mail, which makes it just about impossible to create your own email server without paying out the ass for a whitelist. For social media that is true that bots are harder, instead they make use of grossly underpaid people in 3rd world countries to be human bots, always a way because advertising pays. Kind of why I can't stand advertising in just about any form, its so often underhanded or manipulative.
Oh, yes if its designed with a specific purpose, 4chan was more of an accident on its scale then made with any foresight in mind. Naturally Reddit and Facebook was likely more planned out.