social network

Twitter did not find the influence of “Russian bots” on Brexit. Organized the verification of the British deputy did not accept the answer

It requires further and more detailed investigation – and not the first time.

From interfering in the US presidential elections to influencing the Brekzit referendum

In October 2017, a few days after the Twitter report in front of the US Senate on Russian bots interference in the presidential election, a member of the lower house of parliament and the head of the Committee on Culture, Media and Sport Damien Collins (Damian Collins) addressed the leadership of the social network . In a letter addressed to Twitter CEO Jack Dorsey, he recalled the “American” accounts, in which representatives of the social network found a connection with the Petersburg troll factory “Internet Research Agency”, and asked to report on the same activity of bots in the political life of the UK.

Collins questions in the letter were based on a City University of London reportpublished in October on the study of bots during the Brexit referendum. It was said about almost 13.5 thousand accounts that actively tweeted in the weeks before the referendum in support of exit from the EU, and after voting day were removed (another 26 538 detected accounts changed names and addresses). The bots were about eight times more frequent than ordinary Twitter users wrote slogans in support of exit – such slogans were contained in 31% of their posts (with 17% of posts calling to stay in the EU).

Collins wrote not only on Twitter, but also on Facebook, which also reported to the Senate about fake news on elections. He dubbed his initiative “Investigation of fake news” regarding the UK and set the task of getting a report from the largest social networks on how great Russian bots interfere with British society.

On Twitter, Collins can see how his initiative developed. He followed the statements of Prime Minister Theresa May, who in mid-November accused Russia of interfering in the referendum (but did not specify what it was). This rhetoric was continued by the head of the British Foreign Office, Boris Johnson, who, at a meeting with Russian Foreign Minister Sergei Lavrov in Moscow, caught the Russian minister by distorting his words about Russia’s intervention in Brexit.

Fight against all bots

On November 21, Collins was answered from Facebook, simply promising to sort out and write about the results in three weeks. A more detailed letter from therepresentative of Twitter on November 24 boiled down to the fact that although the company is investigating the activities of bots in relation to the UK, not all bots are bad (as Collins understands it) – there are those who follow the edits on Wikipedia or just publishes data from its Wi-Fi weights to the social network – and there are overly active accounts of living people who resemble bots.

Twitter said that they are responsible for labeling people like “Connected with the Russian trolley factory” on living people and try not to make mistakes, therefore, in some cases they cannot say for sure. Twitter representatives also drew attention to the conclusion of a City University report cited by Collins: the researchers said that although some bots did publish something on the topic, the scale of their activities was so small that it was not able to change the tone of the discussion and what something to influence.

In early December, Collins received the first response on Facebook essentially: there was information only on advertising budgets of those pages that were found to be suspicious during the presidential elections in the United States. The deputy demanded a more detailed report on all accounts, like what Facebook did during the presidential elections in France.

On December 13, he received another letter from Twitter, which stated that the only discovered account that the company had found while studying the financing of advertising on the platform from Russian sources was Twitter of the Russian RT television channel (which Twitter had forbidden to place advertising since October because of influencing US elections). Then Collins called the Twitter response “completely inadequate”, set a deadline for January 18th and demanded detailed answers in an ultimatum – otherwise he promised companies sanctions.

Here are the questions that Collins asked Twitter based on data from City University:

  • Were these accounts real users or bots?
  • Who controlled these accounts?
  • What was the coverage of tweets posted by these accounts?
  • Who deleted the tweets that originated from these accounts?

Answers give rise to new questions.

The Twitter response followed only January 19, but clearly not in the form that Collins demanded: neither the reach nor the account holders were disclosed, and in fact the social network answered only the first question. The main thing is that there was nothing out of the answer that could confirm Russia’s significant attempt to influence the results of the Brekzit referendum.

Twitter representatives said they had requested a database of suspicious accounts from researchers at City University and conducted their own research on it. It turned out that only 1% of these accounts were registered from Russia, 6.5 thousand of discovered accounts were blocked before the report was published, and 44% of the remaining accounts were permanently blocked after studying the City University report.

Another part of the suspicious accounts was not blocked, but they were marked as spam or “low-quality” – Twitter has special algorithms that hide such profiles from the search and exclude their influence on the formation of Twitter trends. The remaining accounts, as it turned out, did not violate the rules of service. In addition, Twitter pointed to a similar study by the Oxford University of the Internet, which studied the bots in Russian Twitter (from 2014 to 2015): scientists found little evidence of the connection of automated accounts actively participating in discussions with “Russian sources”.

Collins was also not satisfied with this answer, especially since on January 19, Twitter published a detailed report on how Russian bots influenced US elections, even citing specific tweets. True, the company was still criticized for insufficient efforts: it sent letters to users who retweeted, liked or subscribed to fake accounts during the US elections, but did not show either these accounts or tweets with which interactions occurred – that is, users could not find out where and how they were tried to be influenced.

On January 25, Collins sent another letter to Jack Dorsey (although all this time he was answered by completely different people). There are more questions.

Twitter claims that 1% of accounts from the University of City published a report […] was registered in Russia. Could you confirm how many other accounts were controlled from agencies in Russia, even if they were not registered there?

How many accounts are similar in characteristics to those already identified as associated with Russia, even if you have yet to prove the existence of such a connection?

Twitter still has to answer whether the 13,493 accounts described in the City University survey, who controlled these accounts, which audience watched their activity during the referendum, and who deleted their tweets were real users or bots.Damien Collinsdeputy of the lower house of parliament of Great Britain and head of the relevant committee

Back to top button