Investigating here there are a bunch of interesting stories.
In the case of:
I can see the user got silenced multiple times by AI.
Somehow we saw the sequence of:
- System flagged
- Flag rejected
- User edited
- Spam system picked it up again
- No flag was created cause it would be a “Duplicate”, so system transparently did nothing.
Let me have a chat with the staff experience team about this to see how we can properly resolve it.
We need to allow AI to re-flag stuff cause spammers can be very sneaky, just cause it was good to start with does not mean it is good forever.