Twitter Improves Filters and ‘Frees’ Name Calling Among Friends

Twitter announced on Wednesday (5) that it has improved its filter of offensive speeches and the warnings that try to make you give up sending posts that contain aggressive language or swearing.

The novelty was initially implemented in 2020 on the social network app for Android and iOS. It consists of a window that appears when you identify certain terms, asking if you really want to send that message and suggesting a review.

According to internal tests of the network, 34% of users gave up sending the tweet if it was selected by the filter, while 11% of the people who resent a message changed the language to make it less offensive.

What has changed?

As of now, starting with English messages only, Twitter has made some changes to the filter based on feedback from the community.

The first is to “release” offensive exchanges of messages depending on the level of user interaction. If the insulting dialogue is between people who follow each other and usually talk to each other without offense, Twitter now understands that the strongest word can be a joke.

In addition, terms reappropriated by marginalized communities and used in a non-offensive manner are also released.

To continue improving the warning, the company added a questionnaire after the process to find out if it made the correct detection of offenses.



Please enter your comment!
Please enter your name here