Design Harmful Content Detection

1. Frame the Problem: What Makes It Its Own Design

"Design a system to detect harmful content." That covers hate speech, harassment, spam, graphic violence, misinformation, the whole list. This is one of the most-asked ML design questions there is, because every platform with a text box eventually needs an answer to it.

It shares the fraud lesson's adversary and its asymmetric costs. But it adds three complications that fraud never had to deal with