Twitter to curb abuse by hiding more bad tweets
In an effort to curb trolling and abuse on the platform, Twitter is making changes to its algorithm to hide more bad tweets in conversations and searches. As part of the algorithm tweak, Twitter will now use a wider range of parameters to rank tweets. This will help the company discern if a tweet is posted by a bot or is in general abusive.
Arranging tweets to influence how often people see them
Once the upcoming changes are put in place, tweets will be ranked on things like the number of accounts a user has, IP address, and if a tweet has got its parent account blocked or muted. Earlier, Twitter only arranged tweets by quality and back-and-forth conversation. Tweets identified to be bad will get cast down into "Show more replies" so fewer people see them.
Twitter trying to identify abusive tweets without human review
Twitter claims that while testing the algorithm change, abuse reports on tweets by users went down by 8% in conversations. With the algorithm tweak, Twitter reduces its dependency on users having to report a tweet as abusive for it to be taken it down. "We want to take the burden of the work off the people receiving the abuse or harassment," Twitter said.
Twitter cleaning your feed of tweets you'd rather not see
With this, Twitter is trying to solve the long-standing problem of abusive tweets on the platform. Content moderation today does need algorithmic intelligence. The company has also promised to communicate how the new enforcement makes decisions and takes action across the platform to keep everything transparent (since algorithms can go wrong too). The changes will roll out globally this week.
