-snip-
I've seen quite a lot of tools wrongly detect a post as AI generated, when it actually isn't. I know you can say it has to pass through several tools before a conclusion can be reached, but if one tool is saying it isn't AI generated and another is saying it is, the accused user has an argument and banning them outrightly might not be possible. Using AI chatbot to post can only attract the same punishment as plagiarism, if it can be completely proven that the post is AI generated without any false positives.
Sometimes what you say is true - however, if an AI detection tool says it's an AI-generated post, then it's most likely true. You can use some detection tools as used by some users in this thread - so this will prevent you from making the wrong decision that someone is posting with AI.
I think @nutildah and @lovesmayfamilis and some other user are pretty good at this. Check out their work in the thread I referenced below:
1.
AI Spam Report Reference Thread