FB Takes First Steps To Curb Fake News, Using AI To Catch Hoax Articles Shared By Spammers

Years after the issue was first dragged into the spotlight during the US elections, Facebook is still fighting fake news, hoaxes, and spammers on its platform. Now, the company announces it¡¯s implementing a few automated measures to help in the war against trolls.

In a new blog post, Facebook announced it¡¯s now using AI to target misleading content, starting with small steps. Of course, machine learning can¡¯t yet do things like fact-check stories or make judgement calls about a piece¡¯s authenticity, but it is good at spotting some red flags.

Facebook already has massive human teams dedicated to stopping fake stories from appearing on the platform, flagging and taking them down as they see them. What the company is now doing is having an AI look through posts for duplicates of stories that have already been debunked.

¡°Machine learning helps us identify duplicates of debunked stories,¡± the company wrote in its post. ¡°For example, a fact-checker in France debunked the claim that you can save a person having a stroke by using a needle to prick their finger and draw blood. This allowed us to identify over 20 domains and over 1,400 links spreading that same claim.¡±

Facebook product manager Tessa Lyons further detailed the system to Buzzfeed News, saying these AI filters are now attempting to predict which pages are most likely to share fake content. This involves looking for page admins that may live in one country but largely post stories about another nationality. They also look for admins that may not be fake accounts but are linked with suspicious activity.

Of course, Lyons admitted it¡¯s possible that even legitimate sites may be accidentally targeted, but she¡¯s confident it should be the minority of cases. And for those pages found to be guilt, Facebook says it will actively reduce their reach and remove their ability to monetise, a huge discouragement for outsourced spammers looking to make a quick buck.

Additionally, Facebook also says it¡¯s now working with third-party fact-checkers in 14 countries now, and plans to scale that up further by the end of the year. After all, AI may not have the capability to do the authenticating of stories humans can, but it can provide some solid backup.