Facebook’s algorithms and A. I. are far from perfect, and an incident yesterday involving stories discussing the company’s recent security breach provide a prime example. The company’s spam filters accidentally stopped users from posting certain stories related to the recent breach.
Yesterday, it was reported that Facebook was hit by yet another security breach. This most-recently discovered vulnerability affected as many as 50 million users. If that security scandal wasn’t bad enough, it has now been reported that Facebook attempted to stop users from sharing certain stories related to the breach.
The issue seems to have affected two stories about the breach. The first is from the Guardian and the second from the Associated Press (AP). The tweet shows a screenshot saying that the story had been blocked due to concerns that it was spam. While this does look bad, recent evidence suggests that this was little more than an unfortunate error on the part of Facebook’s automated spam filters.
Other Facebook users reported that they were not having any trouble with posting the story. Facebook itself has clarified that this was an issue with their automated moderation system. In short, the story was posted so many times in such a short time period that it accidentally triggered the spam filters.
All in all, this explanation makes more sense than some conspiracy on the part of Facebook. After all, news of this breach has been reported by nearly every tech site and has been picked up by mainstream news outlets. However, it does highlight, yet again, the flaws in Facebook’s algorithms and automated software.
Facebook, like pretty much every other tech company on the planet, relies heavily on A. I. to help manage its site. This isn’t necessarily a bad thing since there’s so much content on Facebook that it would be all but impossible for human moderators to review it all. Unfortunately, the timing of this error could hardly come at a worse time for the social media giant. The company is still facing questions regarding data use by Cambridge Analytica, and these recent security issues are not helping matters.
