According to an article published on the website dailymail.co.uk on février 23, 2021 by ,
Twitter is again testing a feature that asks users to revise replies that are potentially harmful. If its AI detects such language, a notification appears with options to revise, delete or send.