I dunno, this sounds a lot like "hey, i'm throwing this net into the ocean and I'm not catching all the fish, booo!" Of course your're not going to stop them all, but the fact that you're stopping many is significant progress. I get that anyone can say anything at any time, even if they are part of the trusted circle, but just having some tools to address many of the cases is much better than nothing. If someone wants to live a 100% jerk-free life, they can put on a straight jacket and walk into a psych ward. (I don't know if they use straight jackets anymore... hahaha).
High rate of false positives is the main problem
Lower the score threshold -> increases false negatives but im okay with that. Low scores could just be new acct. I just want 1) minimize spam (reply bots, nostrichhouse, happy new year spam) 2) accounts that just shotgun toxicity to anyone they interact with i'm okay muting anyone that passes through that filter