Happy Nice Mall

Facebook spokesman Joe Osborne said in a statement that the company was “already investigating these issues” at the time of Allen’s report. “Since then, we have formed teams, developed new policies, and collaborated with our industry peers to address these networks. We have taken aggressive enforcement measures against such foreign and domestic baseless groups and have made the results public quarterly.”

In the process of fact-checking shortly before publishing this story, MIT Technology Review found that five of the troll farm pages mentioned in the report remained active.

The report found that troll farms reached the same demographic groups targeted by the Kremlin-backed Internet Research Agency (IRA) in the 2016 election, targeting Christians, Black Americans and Native Americans. 2018 BuzzFeed News investigation It found that at least one member of the Russian IRA, accused of alleged interference in the 2016 US election, also visited Macedonia at the time the first trawler farms were revealed, but found no concrete evidence of a link. (Facebook said their investigation also did not reveal a link between the IRA and Macedonian troll farms.)

“This is not normal. This is not healthy,” Allen wrote. “We have empowered non-authentic actors to recruit large followings for largely unknown purposes … actors with possible ties to the IRA have access to very large audience numbers in the same demographic groups targeted by the IRA. , the US poses a huge risk for the 2020 election.”

As long as the troll farms were successful in using these tactics, any other bad actor could have succeeded, he continued: “If Troll Farms reaches 30 million US users with content targeting African Americans, the IRA now has a large audience there. ”

Allen wrote the report as the fourth and final part of a year-and-a-half effort to understand trawler farms. According to the former Facebook employee who provided the report, he left the company that same month, in part due to frustration that the leadership had “effectively ignored” his research. Allen declined to comment.

The report highlights the alarming situation in which Facebook leadership has abandoned the platform for years, despite repeated public promises to aggressively combat foreign-sourced election interference. MIT Technology Review is making the full report available, with employee names corrected, because it’s in the public interest.

His revelations include:

• As of October 2019, approximately 15,000 Facebook pages, mostly US ones, were depleted in Kosovo and Macedonia, which were known as bad players in the 2016 election.
• Collectively, these troll farm pages, which the report treats as a single page for comparison, reached 140 million monthly US users and 360 million weekly global users. Walmart’s page reached the second largest U.S. audience with 100 million.
• The troll farm pages also combined to form:
  • The largest Christian American page on Facebook is 20 times larger than the next largest; It reaches 75 million monthly US users, 95% of whom have never followed a page.
  • The largest African-American page on Facebook is three times larger than the next largest; It reaches 30 million monthly US users, 85% of whom have never followed a page.
  • The second largest Native American page on Facebook has reached 400,000 monthly users, 90% of whom have not followed any pages.
  • The fifth largest women’s page on Facebook, reaching 60 million monthly US users, 90% of whom had never followed a page.
• Troll farms primarily affect the US, but also target the UK, Australia, India and Central and South American countries.
• Facebook has conducted numerous studies confirming that content that is more likely to get user engagement (likes, comments, and shares) is of the type that is known to be bad. Still, the company continued to rank content in user’s news feeds based on what would get the most engagement.
• Facebook prohibits pages from publishing only copy-pasted content from other parts of the platform, but does not enforce the policy against known bad actors. This makes it easy for foreign actors who don’t speak the local language to post fully copied content and still reach a large audience. At some point, up to 40% of pageviews on US pages went primarily to those with non-original content or material of limited authenticity.
• Troll farms have previously entered Facebook’s Instant Articles and Ad Breaks affiliate programs designed to help news organizations and other publishers monetize their articles and videos. At one point, 60% of Instant Article readings were going to plagiarized content from elsewhere, due to the lack of basic quality controls. This made it easy for troll farms to mix undetected and even get paid from Facebook.

How is Facebook enabling troll farms and growing their audience?

Specifically, the report looks at troll farms based in Kosovo and Macedonia, which are run by people with little understanding of American politics. Yet due to the way Facebook’s news feed reward systems are designed, they can still have a significant impact on political discourse.