This is gonna be a difficult one probably due to the wide scope and multi-linguistic nature of Commons, but let's start the conversations about what options we have and any blockers there might be.
So, we've been hard at work making multi-lingual contexts work for us. We'll probably need to do some research on how to take advantage of it in the context of commons though.
When https://github.com/wiki-ai/revscoring/pull/206 gets merged, we'll support detection of badwords/informals in the following languages:
It's just a matter of time until we finish the implementations of:
We need a native speaker to help us clean up some automatically generated lists for:
We haven't started work on other languages yet, but we can if there is demand. We'll prioritize working on languages where we can get a native speaker to help us review an automatically generated list of potential badwords.