Quoted By:
When you click that little flag next to the block button it shows you this.
It asks you to say how the message hurts you so it can improve it's AI training.
What I can see could be done is that Chuubas made an habit of reporting when they get insulting marshmallows.
I would suggest that frill that box with both Spanish (For it's AI model) and an machine translation to Japanese just in case their model works better that way, or in case it's manually checked by Japanese staff.
Eventually it should get better at filtering malicious comments.
