LONDON – Twitter Inc., under pressure from governments around the world to combat online extremism, said in a new report that improving automation tools are helping block accounts that promote terrorism and violence.
In the first half of the year, Twitter said it suspended nearly 300,000 accounts globally linked to terrorism. Of those, roughly 95 percent were identified by the company’s spam-fighting automation tools. Meanwhile, Twitter said government data requests continued to increase, and that it provided authorities with data on roughly 3,900 accounts from January to June.
The increasing role of machines in fighting extremism is a function of necessity, with manually identifying violent material within the millions of messages sent every day an impossible task.
Twitter, along with Facebook Inc. and YouTube, are instead building automation tools that quickly spot troublesome content. Facebook has roughly 7,500 people who screen for troublesome videos and posts. It’s also funded groups that produce anti-extremism content that’s circulated on the social network.
Twitter said about 75 percent of the blocked accounts this year were spotted before a single tweet was sent, and that 935,897 accounts had been suspended since August 2015, with two-thirds of those coming in the past year.
“Our anti-spam tools are getting faster, more efficient and smarter in how we take down accounts that violate our policy,” Twitter said in a statement.
The company is balancing a commitment to free speech against pressure from policymakers who want to see social media companies do more to fight extremism and hate speech. While the company is suing the U.S. government in an effort to report more granular information about the national-security requests it receives, Twitter last year signed a voluntary pledge in Europe to take action within 24 hours against reports of racist, xenophobic and violent content.
American authorities made 2,111 requests from Twitter from January to June, the most of the 83 countries tracked by the company. Twitter supplied information on users in 77 percent of the inquiries. Japan made 1,384 requests and the U.K. issued 606 requests. Turkish authorities continued a trend of aggressively policing Twitter, making 554 requests for account data and issuing court orders to remove 715 pieces of content. Other governments made only 38 total content-removal requests.