Skip to main content

Twitter tests hiding direct messages it thinks are abusive

Twitter tests hiding direct messages it thinks are abusive


‘Unwanted messages aren’t fun’

Share this story

The Twitter bird logo in white against a dark background with outlined logos around it and red circles rippling out from it.
Illustration by Alex Castro / The Verge

Twitter is testing a new feature that will automatically hide direct messages that its automated systems think could be offensive. In a short GIF posted by the company, the new section can be seen in the app’s existing “Message requests” folder, which is where Twitter puts any messages sent to you by users you don’t follow. From there, you’d have the option of viewing “additional messages, including those that may contain offensive content,” and Twitter will give you a shortcut to be able to quickly delete any messages that you don’t want to see.

“Unwanted messages aren’t fun,” a tweet announcing the test reads. “So we’re testing a filter in your DM requests to keep those out of sight, out of mind.”

Twitter already has a similar filtering feature for replies to tweets that it introduced back in 2017. Scroll down the replies of any moderately popular tweet, and chances are, you’ll see a small button that allows you to show additional replies that “may contain offensive content.” If the effectiveness of this current tool is any indication, then expect Twitter’s new DM filter to block a lot of messages based on the use of swear words alone since it can already sometimes block otherwise inoffensive replies just because they drop an F-bomb or two.

The new features are part of a broader push by Twitter to reduce the levels of abuse some users suffer on its platform. Earlier this year, it boasted that it’s getting better at automatically detecting abusive tweets rather than needing to rely on them being reported directly, for example. You could argue that it might be more effective to simply ban users who post offensive content before it turns into targeted abuse, but doing so is apparently harder than it sounds.

If Twitter’s latest test becomes a reality, then it could be another good way of cutting down on the amount of abusive content the average person is exposed to. However, Twitter has been testing a lot of new features recently, and few of them have been released. Here’s hoping this latest feature doesn’t get trapped in this testing phase.