Foreign efforts to influence 2018 US elections on Twitter 'limited': company

Twitter Inc took down nearly 6,000 tweets during last November’s U.S. congressional elections, most of which were attempting to suppress the vote through intimidation or sharing false information, the company said on Thursday.

Most disinformation around the elections originated inside the United States, the company said, though it also identified “limited operations” to influence voters potentially connected to sources in Iran, Venezuela and Russia.

U.S. intelligence agencies concluded that Russia ran an interference operation of hacking political party computers and using social media during the 2016 U.S. presidential election campaign to undermined the American democratic process. Moscow denies any meddling.

Since then Twitter, Facebook Inc and other social media companies have been under pressure globally to remove propaganda, bots and illegal content. The companies have stepped up controls, but are still criticized by lawmakers and others for being too slow to close rogue accounts.

Twitter said it identified fewer foreign influence attempts than in 2016 and suspended a majority of the abusive accounts before the elections for the U.S. Senate and the House of Representatives.

An additional 418 accounts possibly affiliated with Russia’s Internet Research Agency (IRA), which U.S. prosecutors said spearheaded the 2016 influence operation, were identified by Twitter. It previously disclosed 3,843 accounts in that category.

“Most appear to originate in Russia, and much of the behavior mimics the activity of prior accounts tied to the IRA,” the company said, but noted that it could not confirm a definitive link.

The company also said it had removed 1,196 accounts in Venezuela that “appear to be engaged in a state-backed influence campaign targeting domestic audiences,” and suspended 2,617 malicious accounts believed to originate in Iran. The Venezuelan accounts were taken down in December 2018, Twitter said.

The company examines 8 million accounts per week for abusive or inauthentic behavior and ultimately removes about three-quarters of them from the service, company representatives said.

Reporting by Katie Paul; editing by Grant McCool