Twitter has announced that it is rolling out a new feature for all English-speaking iOS and Android users that detects “potentially dangerous or offensive” replies and directs people to review their tweet before hitting the send button. Social media platforms started testing this request exactly one year ago prompting people to “pause and reconsider” replies containing “insults, offensive language, or hate comments” before vomiting them up on Twitter. Once prompted, people can then edit their tweet, delete it immediately, or send an as-is reply. But those initial tests sometimes render people unnecessary as algorithms struggle to pick up on nuances in everyday conversation and “often fail to distinguish between potentially offensive language, sarcasm and friendly banter,” Twitter said. Since then, Twitter has made improvements to the system that decides when and how these reminders are sent based on public feedback and lessons learned from the test. For example, algorithms now factor into the nature of the relationship between writers and commenters because people who follow each other are more likely to make fun of than strangers. The system also takes a situation-oriented approach in cases where language can be reclaimed by underrepresented communities and used in a way that is not dangerous.
