Most people would like to access the web without worrying about content they find upsetting. But text classification is very hard to get right.
Identifying hate speech is a difficult text classification problem.
Facebook has issues with
Somewhat successful classification of news sites into fake/reliable