From bots to George Soros: Facebook, Twitter and Russian interference

11 Sep 2017

Facebook HQ at Menlo Park, California. Image: achinthamb/Shutterstock

More evidence of Russian interference during 2016’s US presidential election is now materialising.

On Wednesday 6 September, Alex Stamos, chief security officer at Facebook, confirmed in a blog post that approximately $100,000 in ad spending – from June 2015 to May 2017 (roughly 3,000 ads) – was carried out by 470 inauthentic accounts that likely operated out of Russia.

As The Verge pointed out, Facebook’s official standpoint until then was that it was unaware of any cases of Russia-based individuals or groups buying ads on the site. This was the company position as recently as 20 July, with a statement reading: “We have seen no evidence that Russian actors bought ads on Facebook in connection with the election.”

An army of bots

Melvin Redick appeared to be a man from Harrisburg, Pennsylvania, posting a link on Facebook to a website called DCLeaks. The website contained stolen material from people such as George Soros and higher-ups in the Democratic party – material that would influence the election proceedings.

Redick, The New York Times found, was a fake name, one of many used by Russian operatives during the election process. In an investigation with cybersecurity firm FireEye, the paper detailed some of the mechanisms used by these operatives to spread contentious propaganda.

Most of the accounts on Facebook were not explicitly anti-Clinton, according to Stamos: “Rather, the ads and accounts appeared to focus on amplifying divisive social and political messages across the ideological spectrum; touching on topics from LGBT matters to race issues to immigration to gun rights.”

According to The New York Times, there were hundreds of Twitter accounts posting anti-Clinton messages and drawing attention to leaked material from Russian operatives. Accounts such as DCLeaks, Guccifer 2.0 and Anonymous Poland were found to be associated with each other, FireEye concluded.

Lee Foster of FireEye suggested that some of these accounts may have been purchased on the black market, while accounts of unsuspecting individuals were simply hijacked.

Possible Russian motives

According to Anne Applebaum of The Washington Post, although the content of the emails revealed by DCLeaks were “utterly banal”, the use of anti-Clinton hashtags on Twitter, as well as narratives, inflammatory statements and emphasis on liberal financier George Soros, persuaded people to click.

In terms of the motivations behind Russia’s involvement, she wrote: “In backing Donald Trump, Russia’s oligarchical class sought not only to disrupt US politics but also to reverse sanctions, both those applied in the wake of Russia’s invasion of Ukraine and those connected to the Magnitsky Act, which targeted officials involved in human rights violations.”

High alert

Companies such as Facebook and Twitter will now have plenty of food for thought and considerations to make ahead of the 2018 congressional vote and 2020’s presidential election.

Former FBI agent Clinton Watts described the situation on Facebook and Twitter to The New York Times as a “bot cancer eroding trust on their platforms”.

He believes that Facebook is currently doing more to tackle the issue than Twitter. Twitter itself made a statement in June 2017, saying: “We cannot distinguish whether every single Tweet from every person is truthful or not. We, as a company, should not be the arbiter of truth.”

In the aforementioned blog post, Facebook’s Stamos emphasised his company’s efforts to improve its systems around detecting false or malicious profiles: “Over the past few months, we have taken action against fake accounts in France, Germany and other countries, and we recently stated that we will no longer allow Pages that repeatedly share false news to advertise on Facebook.”

“Along with these actions, we are exploring several new improvements to our systems for keeping inauthentic accounts and activity off our platform. For example, we are looking at how we can apply the techniques we developed for detecting fake accounts to better detect inauthentic Pages and the ads they may run.

“We are also experimenting with changes to help us more efficiently detect and stop inauthentic accounts at the time they are being created.”

How far a reach?

While the DCLeaks and propaganda attempts by Russian operatives did contribute to the divisive nature of the election, it is hard to find figures from Facebook as it is not releasing any of the Russian ads it found. Operatives apparently used Facebook’s ‘dark ads’ in these operations.

This type of ad can be shown to individual users by an advertiser without leaving a trace, and disappear as users scroll by.

The Daily Beast estimated that the adverts could have reached up to 70m people, but more details on Facebook’s targeting and algorithms will be the only real way to find out.

Facebook HQ at Menlo Park, California. Image: achinthamb/Shutterstock

Ellen Tannam was a journalist with Silicon Republic, covering all manner of business and tech subjects