Twitter tests filter of potentially abusive messages

Twitter Team Announces Testing Filter Of Messages Containing Offensive Content. Screened out

in this way letters will be sent to a separate folder.

How it works

In 2019, Twitter began to take the firststeps to identify and block offensive tweets without human intervention. However, the system does not work perfectly: there is still a lot of dirt in the social network, and false positives are inherent in the algorithm.

The developers found that most of the "Chernukha"accounts for personal messages from completely strangers, especially various minorities and women. A new filter should fix this problem. With it, all potentially offensive letters will be sent to the "additional messages" folder. At the same time, the first few lines will be overlaid with the note “This message is hidden because it may contain offensive content.” Further, it all depends on the user whether he will be interested to read such a message or not.