Social media companies are keenly focused on ridding their platforms of hateful speech and harassment. But content moderation can only capture a small fraction of it, and automatic filters can risk censoring ordinary users. Another approach is to counter these messages with opposing points. IPL researchers tested various kinds of counterspeech on Twitter, and they found that messages invoking empathy worked better than any other—they led authors of hate speech to discontinue their posts and even delete old ones.
