to quote something I said a day ago about AI spotting in the posts of other people:
https://news.ycombinator.com/item?id=46114083
"I think that writing style is more LinkedIn than LLM, the style of people who might get slapped down if they wrote something individual.
Much of the world has agreed to sound like machines."
This has always been the case in the "corporate/professional" world imo.
It's just much easier now for "laypeople" to also adjust their style to this. My prediction is people will get quickly tired of it (as evidenced by your comment)
AI witch-hunts are definitely a problem. The only tell you can actually rely on is when the AI says something so incredibly stupid that it not only fails to understand what it is talking about but the very meaning of words themselves.
Eg,metaphors that make no sense or fail to contribute any meaningful insight or extrenely cliched phrases ("it was a dark and stormy night...") used seriously rather than for self-deprecating humor.
My favorite example of an AI tell was a youtube video about serial killers i was listening to for background noise which started one of its sentences with "but what at first seemed to be an innocent night of harmless serial murder quickly turned to something sinister."