Facebook to use machine learning to fight engagement bait

Facebook to use machine learning to fight engagement bait 1

Facebook, one of the largest social media, which is used by appx 2 billion active users monthly (According to wikipedia), decided to use machine learning to fight Engagement Bait. It is common on the facebook platform that user request or tag some users to increase user engagement on their posts and sometimes that seems to be annoying.

Facebook was getting lots of complaint against it and finally facebook decided to make a move to control it using technology.

There are many posts on Facebook that force users to like or share using some emotional and sentimental messages like “like if you are Aries” or “Share if you love your mother” etc. that is called engagement bait and Facebook start demoting individual posts from those people and Pages.

Facebook

Facebook has categorised hundred of thousands of posts to inform a machine learning model that is capable of detect these kind of posts and start demoting on it. In coming month Facebook is planning to show stricter demotion on pages that are repeat offenders. According to facebook —

Additionally, over the coming weeks, we will begin implementing stricter demotions for Pages that systematically and repeatedly use engagement bait to artificially gain reach in News Feed. We will roll out this Page-level demotion over the course of several weeks to give publishers time to adapt and avoid inadvertently using engagement bait in their posts. Moving forward, we will continue to find ways to improve and scale our efforts to reduce engagement bait.

Engagement Bait

Facebook has cleared one thing that posts with Missing child information, Asking for help,  advice, or recommendations, raising money for a cause, or asking for travel tips etc will not be impacted by this update.

Facebook has also updated its newsfeed guideline for users so that publisher, businesses or any user can be sure of not being the involved into it and prevent them self of being demoted.