Here are a couple of posts for reference, but I think the general consensus is to use AI to handle this. You could also have a list of inappropriate words and built your own check, but it’s harder to prevent false positives or missed words.