google.com, pub-5232289331708098, DIRECT, f08c47fec0942fa0
News

Twitter Says A.I. Is Now Removing Over Half of Its Abusive Tweets Before They’re Flagged

Tackling abuse on Twitter has been an ongoing problem for the social network, but it’s one that its automated moderation tools are apparently getting better at handling.

In a letter to shareholders on Thursday, Twitter reported that 50% of all abusive tweets on the platform are being removed by its automated moderation tools, before users have a chance to report them. The statistic only includes abusive tweets that have been removed from the site and does not count controversial tweets that have been allowed to stay up, or abusive tweets that have yet to be flagged.

Twitter uses a machine learning model that appears to be getting better at detecting abusive tweets and quickly removing them before they’re seen by many users. Last quarter, Twitter’s automated moderation tools were responsible for removing 43% of the total tweets that were deemed in violation of Twitter’s policies. The quarter before that, Twitter’s systems were catching 38% of abusive tweets, while relying on users to flag the rest.

Twitter has also had to walk a fine line when dealing with tweets from world leaders that could be seen by some as abusive. In June, the company announced it would allow tweets that are of public interest to remain up, but would add a warning label and explanation for why it is allowed to stay up, before a person is able to view it.

Tackling abuse has been key for Twitter as the company seeks to grow its daily active user base, a key component to driving revenue and appeasing Wall Street.

In July 2018, Twitter began purging suspicious and dormant accounts from the site, which gave many people an inflated follower count. The move was met with criticism from President Donald Trump, who lost around 300,000 followers in the purge.

More than one year later, Twitter’s user base is growing, as the company continues to weed out automated accounts.

Twitter added 6 million new users in the third quarter, putting its total daily active user base at 145 million people, according to the latest earnings report. That’s 21 million more users than this time last year. The positive user and abuse numbers were not enough to lift Twitter’s stock performance today, however, as the company reported that bugs in its advertising products caused it to miss, revenue estimates, driving its shares down 20%.

More must-read stories from Fortune:

—The wireless industry needs more airwaves, but it’s going to be costly
—Demand for Apple’s new iPhone 11 is strong
—How to claim a cash settlement of up to $358 for Yahoo’s data breaches
—Now hiring: people who can translate data into stories and actions
—Investors are pouring money into marijuana software. Here’s the latest startup to get funding
Catch up with Data Sheet, Fortune’s daily digest on the business of tech.

Source link

Comments Off on Twitter Says A.I. Is Now Removing Over Half of Its Abusive Tweets Before They’re Flagged