Artificial intelligence is being developed that will allow advisory "quarantining" of hate speech in a manner akin to malware filters - offering users a way to control exposure to "hateful content" without resorting to censorship.