I just stumbled upon this article on Wikipedia about identifying sings of AI writing. It is specifically about Wikipedia articles but some can be generally applied to other texts as well. For example the Ko-fi spam wave we had once a few months ago checked many of the boxes like overuse of em-dashes, emoji bullet points, a lot of bold text and title cases in headings (I bet most forum users don’t even know you can make headings here).
So I post this because I think it is still useful for us as a comunity even though it is meant for Wikipedia editors.
Other signs noticed while moderating and having to deal with spam a lot on the forum is that AIs tend to switch subjects abruptly and struggle to stay on topic even in a single post. I guess because their ultimate goal is to get users to click on a spam link at the end of the post or even make them search a term or brand or product themselves that was then mentioned in the post.
