Facebook Blocked Legit COVID-19 Posts By Mistake
Given how it has deeply suffered and continues to suffer for misinformation and spam, Facebook has implemented more stringent measures to make sure its platform remains in the good graces of users and of governments. It seems, however, that it may have gone overboard either through oversight or technical error. A few Facebook users have reported that their posts were blocked because of spam, even when those links pointed to official and authoritative information on the COVID-19 coronavirus.
Facebook has stepped up its efforts to combat spam and "fake news" from proliferating on its social platform. It has employed both human moderators as well as AI-based systems to filter out links and content that violate its guidelines. It seems that the latter is to blame for an overzealous culling of links, including but not limited to the ongoing pandemic.
Facebook announced that it was sending home its content moderators as part of its efforts to protect its employees. It is, therefore, relying mostly on its machine learning systems to review posts and block harmful or misleading content. The company warned that mistakes might happen and this latest puzzling incident could have been one of them.
Facebook VP for safety and integrity Guy Rosen more or less confirms that the automated system was at fault. To dispel any conspiracy theories, the exec explained that even posts not related to COVID-19 were removed by mistake. All of those have now been restored.
We've restored all the posts that were incorrectly removed, which included posts on all topics – not just those related to COVID-19. This was an issue with an automated system that removes links to abusive websites, but incorrectly removed a lot of other posts too.
— Guy Rosen (@guyro) March 18, 2020
The brief incident highlights one of the disadvantages of relying completely on automated systems for such tasks. At the same time, the current situation with COVID-19 quarantines does provide Facebook the stress test it needs to see how well or, in this case, how poorly the automated system is when left to its own. Fortunately, it was at least able to get to the root of the problem at this critical period when social media is both at its best and at its worst.