Researchers have developed a new method for automatically detecting hate speech on social media using a Multi-task Learning (MTL) model, they discovered that right-leaning political figures fuel online hate