< Return to blog

A new analysis suggests that Facebook has not been able to keep up with its promises for efficient content moderation policies

A new analysis suggests that Facebook has not been able to keep up with its promises for efficient content moderation policies

Not too long ago, the Wall Road Journal reported Fb’s tardiness within the enforcement of insurance policies to take away hateful and dangerous content material from its platform regardless of all its latest guarantees. Plainly both Fb doesn’t care concerning the magnitude of this misinformation and doesn’t wish to cease it, or it’s unable to implement higher insurance policies and methods to maintain this challenge.

Fb solely begins speaking about it when it faces intense exterior strain, just like the latest #StopHateForProfit marketing campaign and the advertiser spend block that it confronted, or earlier incriminating incidents too. The corporate solely tightens its reigns round content material moderation for a short while after going through critical backlash, but it surely doesn’t final lengthy, and shortly one other piece of reports springs up incriminating Fb another time.

Lately, Fb confronted varied challenges, however 2020 was particularly robust, with a bulk load of misinformation across the coronavirus pandemic, heated discussions and disinformation about local weather, and now transparency and safety points within the upcoming US Normal elections. Fb has lots of content material on its plate to reasonable, and its AI fashions and machine-learning methods aren’t precisely working as effectively as they need to have been, primarily specializing in eradicating the content material that has the very best probabilities to go viral.

This was lately analyzed by the Wall Road Journal. They reported round 276 poisonous content material items to Fb in September. These content material items revolved round violence, hatred, and disinformation that was harmful if went viral. Fb’s content material moderation methods took down solely 32 out of those 276 instances. When the Wall Road Journal inquired concerning the remaining instances, that was when Fb realized and confirmed that fifty% of these remaining posts ought to have been eliminated immediately. They did take away that fifty% of content material after 24 hours although, but it surely nonetheless was not fast or environment friendly sufficient as a result of many different posts from the identical set have been eliminated after two weeks! It exhibits that Fb’s content material moderation insurance policies and methods are nonetheless missing regardless of a lot criticism from in all places.

Different customers can also report such content material that violates the platform’s tips, however it’s nonetheless a questionable matter whether or not the corporate can pay heed to these experiences or not?

When the Wall Road Journal’s latest evaluation surfaced, Fb’s spokesperson Sarah Pollack referred to as the evaluation as unreflective of the general accuracy of the corporate’s posts assessment methods. She additionally mentioned that Fb has began relying fairly closely on AI methods amidst the coronavirus pandemic. However which means Fb’s AI methods aren’t working effectively too. They might have been eradicating extra posts as an alternative of much less! So, this reasoning by Sarah Pollack does not likely add up.

The Wall Road Journal reported in August about Fb’s refusal to implement and enhance hate speech insurance policies after an incident in India however seems like Fb has not discovered its lesson, and it doesn’t even appear possible that it’ll be taught it and mend its methods anytime quickly.

A version of this post was first published at
Read More at ----

Ready to get started?