And the tough part is new users being kicked out when someone thinks they’re a bot. How do we protect people from the bot witch-hunt?
The key feature of witch hunts as a concept is the non-existence of witches. If you have a hunt for something that certainly exists (Soviet spies infiltrating US institutions come to mind) it definitively isn't a witch-hunt.
For my part, I'm going to keep downvoting comments that say nothing in the most bland way imaginable, and flagging anyone who posts "I put your question into ChatGPT; here's what it says" (in contexts other than discussions of chat bots).
Related: Reverse Turing Test: https://en.wikipedia.org/wiki/Reverse_Turing_test
New users can't participate unless they go out of their way to find something they might not be interested in and comment until they have enough points to participate elsewhere.