What if we could use rules, tests, and parameters to isolate hate speech? Can we identify and analyze elements like speaker intent, context, identity, tone, audience, or any number of indicators that transform words into meanings and change an innocuous statement into a verbal assault? Combating the proliferation of online hate speech and understanding its mechanics is a complex undertaking

We believe, however, that it can be done. And one way we are working to do so is by teaching machines to recognize hate..