Facebook is once more to clarify how stories spread through the news feed, this time after posts that are "sensational and provocative." The goal is not just to cut down on clickbait, but to cut down incorrect and problematic posts that do not completely guarantee a direct ban on the site.
In a blog post, Facebook manager Mark Zuckerberg writes that people naturally engage themselves more with sensational content. Engagement with this content, he says, increasing the closer it becomes so problematic that it must be banned.
So, instead of moving the line to what is forbidden, Facebook will change distribution algorithms. Posts like Facebook's AI detect as unnecessary provocative, will be distributed less and less, preventing them from seeing a spike in engagement.
Zuckerberg believes that this will put people off creating and publishing this kind of content initially, which ultimately leads to a better experience for users and less polarization.
Facebook uses the term "borderline content" to describe which stories will be affected here. The blog post provides two examples: "Images near the line of nudity, like revealing clothes or sexually suggestive posts" and "posts that do not fall under our definition of hate speech, but which are still offensive." Zuckerberg simply states that the algorithm is adjusted in terms of nudity, but the implication is that it concerns news and more.
"Divisive groups and pages" will also be suggested to users less often, as a result of these changes. 1
Here is a completely non-scientific diagram Facebook designed to illustrate how this works:
I appreciate the sketchiness of the arrow.
The changes were announced today along with a host of other updates on how Facebook handles problematic content. Facebook said it would expand its complaint process for content diet; The company will also create an independent supervisory board intended primarily to act as the Supreme Court in dealing with challenging moderation issues, as a way to partially solve the problem from Facebook's hands.