How Twitter Is Stepping Up Its War Against Online Abuse

July 20, 2017, 6:27 PM UTC

Twitter has spent the past several months stepping up its efforts to curb instances of online abuse and harassment on the social media service. On Thursday, the company released new internal data showing some of the progress it’s made in cutting down on some users’ bad behavior.

Twitter said it is “taking action” against 10 times more abusive accounts than it was at the same time last year, which means the company is levying account suspensions—or, limiting users’ ability to use the service—on “thousands more abusive accounts” per day. Twitter also said it has removed twice as many repeat offenders—users who create new accounts after receiving a suspension—over the past four months.

Perhaps more notable is that Twitter claims it is “driving change in behavior” among users who have been punished for abusive actions on the service in the past. The company said it’s seen a 25% dip in abuse reports for accounts that were previously disciplined, and 65% of users who have been disciplined do not repeat their abusive actions. Twitter added that its new quality filter—which allows users to mute certain keywords and filter out notifications from bot accounts—has led to the company seeing a 40% drop in the number of blocks issued by accounts that have received an “@mention” from accounts that don’t follow them.

“While there is still much work to be done, people are experiencing significantly less abuse on Twitter today than they were six months ago,” Ed Ho, Twitter’s consumer product and engineering general manager, wrote in a blog post on Thursday.

Over the past year, Twitter has debuted new features meant to reduce the prevalence of online vitriol on the social media service, including using an algorithm to detect abusive behavior before it is reported so that Twitter can reduce the visibility of such posts and take action against the perpetrators. That’s part of Twitter’s efforts to temporarily limit the functionality of offending users’ accounts, and the company says it also explains its reasoning to users who it disciplines in the hopes of preventing further harassment.

Get Data Sheet, Fortune’s technology newsletter.

Twitter has often boasted about its commitment to the principles of free speech, sometimes leading to criticism from those who felt the company was more interested in tolerating all forms of speech than in trying to ensure a harassment-free experience for its users.

The presence of harassment on Twitter especially came under the spotlight during last year’s presidential election, as political divides seem to have resulted in an uptick in vitriol and abuse on the service. There have even been calls for Twitter to ban the account of President Donald Trump, an avid Tweeter who has been accused of promoting violence and abuse with some of his posts on the site. (This week, Twitter told Recode that the company holds the president to the same standards as other users. Twitter’s vice president of Trust and Safety, Del Harvey, said: “The rules are the rules, we enforce them the same way for everybody.”)

Meanwhile, the ability to effectively police abusive behavior could prove essential to the company’s financial health, as Twitter needs to show regular, significant user growth in order to appease investors who are concerned that the company is lagging behind social media rivals like Facebook and Instagram. Twitter’s shares received a boost earlier this year when the company reported its biggest quarterly increase in monthly users since 2015, but the company needs to keep repeating that success, and maintaining a positive (and abuse-free) user experience is one piece of the puzzle (along with streaming video).

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward