Twitter is taking the first real proactive step to combat abusive content on the site.

Anil Dash noticed a new feature when he logged into Twitter on Monday called “quality filtering.” It looks like Twitter is finally attempting to filter out tweets containing threatening language.

“Quality filtering aims to remove all Tweets from your notifications timeline that contain threats, offensive or abusive language, duplicate content, or are sent from suspicious accounts,” says the description for the new quality filter.

Well, that's an interesting & welcome addition, Twitter! (Was prompted about this on opening the app.) pic.twitter.com/Ka2VDvqwNf

— Anil Dash (@anildash) March 23, 2015

Twitter later confirmed that this was something it’s testing, but only for verified users at the moment. It’s kind of similar to the “tailored notifications” system that lets users choose to only receive notifications “tailored for you” – but obviously this is more aggressive as it attempts to weed out any sort of harassing tweet.

Twitter has been making steps to address the prevalence of abusive content on the site – but up until now those steps have all been reactionary. Twitter tripled the amount of people handling abuse reports. It made it easier to report abusive content. It banned revenge porn. Twitter tried to make the process of filing a police report simpler. All of these things help, in small ways, but are all based on the idea of user reporting.

This new notifications filtering is Twitter trying to make sure the abusive content never gets to you. Sure, it’ll still be out there in Twitterland – but it won’t pop up in your notifications as an @ mention.

When Twitter CEO Dick Costolo publicly acknowledged his company had a major problem with abusive content, he floated that the solution had to do with visibility. Basically, people can say what they want but Twitter doesn’t have to give them a megaphone.

It looks like this new feature follows that logic. It’s yet to be seen how well it’ll work, and whether the filter will be purely algorithmic or have some sort of human involvement. It’s also unknown when this new feature will be available for average Twitter users. I’ve reached out to Twitter for some clarification.

Image via Rosaura Ochoa, Flickr Creative Commons