‘Massive ranking failure’ meant Facebook showed users nudity, violence, and Russian misinformation

Misinformation and other harmful posts rose in visibility up to 30 per cent globally because of the failure

(Getty Images)

A “massive ranking failure” of Facebook’s news feed meant that half of all views were exposed to “integrity risks” over the past six months.

A new report from The Verge, based in internal documents from Facebook, suggests that since October 2021 a huge amount of misinformation began surfacing on the news feed. Instead of suppressing accounts that repeatedly shared misinformation, it was instead giving them a boost – up to 30 per cent globally, it is claimed.

The engineers were unable to find the cause of the issue, and misinformation surged on the site until 11 March despite it being a high-priority case, according to the report.

During this time, Facebook also allegedly failed to properly moderate nudity, violence, and Russian media. This is despite the company “prohibiting Russian state media from running ads or monetising on our platform anywhere in the world”, firm’s head of security policy, Nathaniel Gleicher, said late February. In response, Russia marked Meta, the parent company of Facebook and Instagram, as an ‘extremist’ organisation.

A spokesperson for the company said that it “detected inconsistencies in downranking on five separate occasions, which correlated with small, temporary increases to internal metrics.” The issue was apparently first introduced two years before it made a noticeable impact on the site.

“We traced the root cause to a software bug and applied needed fixes,” the spokesperson said, but the big “has not had any meaningful, long-term impact on our metrics”.

Facebook posts which do not violate its “Community Guidelines”, and so avoid being banned entirely, but are still deemed to be low quality are hidden within the feed.

The categories of posts that will trigger those demotions are detailed in a document the company released in September last year, known as Content Distribution Guidelines.

They include posts that have spam, that have false information, and which put people at personal risk.

“We’ll continue to update the Content Distribution Guidelines to provide people with information about how we define and treat problematic or low-quality content that doesn’t otherwise violate our Community Standards,” Facebook said.