Twitter says automated tools now flag 38 percent of abusive tweets

Twitter says automated tools now flag 38 percent of abusive tweets

Twitter said its automated systems proactively flag 38 percent of abusive tweets removed from the platform, touting the figure in an announcement on Tuesday.

Just a year ago, that number was zero percent, as Twitter's policies dictated that it would only take action against tweets reported by users. 


Twitter faced widespread criticism for that policy, with critics claiming it was a method of abdicating responsibility over the platform's rampant harassment problem. Over the past year, Twitter has invested in systems to detect "abusive" content before it is reported to them.

"People who don’t feel safe on Twitter shouldn’t be burdened to report abuse to us," Twitter wrote in a blog post on Tuesday. "Previously, we only reviewed potentially abusive Tweets if they were reported to us. We know that’s not acceptable, so earlier this year we made it a priority to take a proactive approach to abuse in addition to relying on people’s reports." 

Twitter's Tuesday blog post highlights multiple statistics that it says reveal the progress the company has made when it comes to the "health" of conversation on the platform. "Health" refers to Twitter's ongoing efforts to root out content that is hateful, harassing, untrue or otherwise inflammatory. 

Content described as abusive by Twitter includes posts that are deemed hateful or threatening. 

The platform has long faced criticism that it incentivizes outrageous or cruel behavior, leading to a polarized online environment. 

Twitter on Tuesday said it suspended 100,000 accounts that users who had been suspended earlier this year created to get around their suspensions.

It also said it has been acting more quickly in responding to requests from users appealing decisions to take down content, and that it has been able to suspend three times more abusive accounts within 24 hours after a report compared to last April. 

Twitter in the post said it is continuing to improve the technology that flags abusive content before it is reported. 

It also announced that, starting in June, it will allow Twitter users to hide replies to their tweets.

Updated at 2:52 p.m.