Twitter Rethinks Abusive Comment Policing

Twitter is cracking down on abusive and personal attacks. It's giving users extra controls while introducing automated checks to pick up on potential abuse.

The new rules aren't specifically about what people can post in a normal tweet. Instead, they cover posts where the writer deliberately includes somebody else's user name, meaning that other person will see the post as part of their notifications.

That helps deal with a loophole in the way Twitter works. For the most part, people only see posts from accounts they've chosen to follow, but the notifications tool includes comments posted by anyone that are intentionally aimed at the person. The only way around this is to completely switch off notifications from people you don't follow, which could mean missing out on positive or helpful communications and making new connections.

Algorithm Aims To Catch Offenders

Until now, the system has largely relied on people who are the target of abusive comments reporting them to Twitter. Some victims of sustained campaigns of abuse have said this simply isn't practical if they come under fire from a large number of people working together.

Instead, Twitter will now attempt to use algorithms to proactively discover abusive posters. One example it gave is if a user repeatedly posts tweets aimed at someone else who doesn't follow them on the site and has not otherwise previously engaged with them. (Source: buzzfeed.com)

Twitter says that they "aim to only act on accounts when we're confident, based on our algorithms, that their behavior is abusive." That implies it might automatically punish people without manually checking what they've written - something that could be controversial if the algorithm is mistaken. (Source: twitter.com)

Anonymous Comments Can Be Blocked

People identified as having posted abusive comments in this way won't automatically be banned. Instead, there will be lighter punishments such as a temporary period during which the only people who can see the user's posts are those who have actively chosen to follow them.

Another set of changes will affect all users. They will now be able to choose not to see any notifications from users missing particular pieces of account information such as a profile photo, a verified email address or a phone number. Such accounts are more typically associated with users who want to post abusive comments anonymously.

What's Your Opinion?

Do you use Twitter, and if so do you find abusive comments a problem? Is an algorithm the way to go considering the scale of the task, or does every case require human moderation? Are the changes sufficient or does Twitter need to act further?