Exclusive: Twitter deletes more than 10,000 accounts that sought to discourage US voters

WASHINGTON (Reuters) – Twitter Inc. (TWTR.N) removed more than 10,000 automatic accounts with messages that dissuaded voters from voting in Tuesday's US election, and mistakenly appeared to belong to Democrats, after the party reported misleading tweets to the social media company.

"We acted on accounts and relevant activities on Twitter," said a spokesman for Twitter in an email. The moves took place in late September and early October.

Twitter has removed more than 10,000 accounts, according to three sources close to the effort of the Democrats. The number is modest, given that Twitter has already removed millions of accounts that he said were responsible for spreading false information during the 2016 US presidential election.

Yet the referrals represent a quick victory for a nascent effort by the congressional congressional campaign committee, or DCCC, a party that supports Democrats who run in the House of Representatives.

The DCCC launched this initiative this year in response to the party's inability to respond to millions of accounts on Twitter and other social networking platforms spreading negative and false information about the Democratic presidential candidate Hillary Clinton and other party candidates in 2016. Operation told Reuters.

While the prevalence of disinformation campaigns has so far been modest in the run-up to the Nov. 6 congressional elections, Democrats hope that the signaling operation will help them react quickly if such messages come up in the next days.

Tweets include those that discouraged Democratic men from voting, saying it would silence women's voices, according to two sources close to the signaling operation.

The DCCC has developed its own system for identifying and reporting malicious automated accounts on social media, according to sources from all three parties.

The system was built in part from publicly available tools called "Hoaxley" and "Botometer" developed by computer scientists at Indiana University. They allow a user to identify automated accounts, also called bots, and analyze how they disseminate information on specific topics.

"We made Hoaxley and Botometer free because they deserve to know what is a bot or not," said Filippo Menczer, a professor of computer science and computer science at the University of Indiana.

The National Democratic Committee works with a group of contractors and partners to quickly identify misinformation campaigns.

Among them, RoBhat Labs, a company whose website says it has developed a technology that can detect robots and identify political biases in the messages.

The collaboration with RoBhat has already led to the discovery of malicious accounts and publications, which have been referred to social media companies and other campaign leaders, said Raffi Krikorian, DNC's director of technology in a statement. email.

Krikorian did not specify whether the marked publications had finally been deleted by Twitter.

"We provide the DNC with reports on what we see in terms of bot activity and amplification," said Ash Bhat, co-founder of RoBhat Labs.

"We can not tell you who is behind these different operations, Twitter is hiding it from us, but with the technology you know, when and how it happens," said Bhat.