By Carolina Martínez Elebi, associate researcher at OBSERVACOM.
Meta’s updated community standards for its social media platforms on January 1 detail how the company applies “visibility reduction” to certain content it considers “problematic” or “low quality,” which neither directly violates its policies nor is reviewed by humans. The vagueness of the measure’s scope is aggravated by the fact that it is discretionary in its application and entails the exclusion of certain expressions from the reach of other users.
In recent months, Meta has not only been one of the tech companies denounced by various users for applying “shadowbanning” measures on pro-Palestinian content, but the company has also been fined more than €27,000 for carrying out the same on the Facebook account of MEP Tom Vandendriessche, “without clear justification”. In this case, the Belgian Court of Appeal found that the lack of transparency and the impossibility of appealing the measure violated the General Data Protection Regulation (GDPR), significantly affecting the politician’s ability to disseminate his expressions.
But what does Meta say about this measure of reducing scope that is becoming more widespread every day across different platforms in a way that could constitute “hidden censorship”?
In the “Taking Action” section of its Transparency Center, the company explains that when applying content rules on Facebook and Instagram, since 2016 they have adopted a three-pillar approach: suppress, reduce and inform.
While the “suppress” posts pillar is the most easily recognized and recognizable—as it involves classic content moderation actions like removing content and deleting user accounts— the “reduce” pillar is explained like this : “There is some problematic content that can create negative experiences for people on Facebook and Instagram. We tend to reduce the distribution of this content, even when it is not fully covered by the removal guidelines in our policies.”
In the section on its “reducing the distribution of problematic content” policy, the company explains that this applies to content that it considers “problematic or low-quality,” even if it does not violate the Community Standards which have been unified and apply to Facebook, Instagram, Messenger, and Threads. This includes posts with titles that function as clickbait, engagement bait (a tactic that encourages interactions to increase the reach of a post), links to websites saturated with ads, repetitive or copied comments, and “content with limited originality.” It also covers cases such as misinformation or creators who repeatedly violate policies.
One thing to note is that Meta says that in such cases, it “may reduce its distribution, consistent with user controls” using the “sensitive content” control within the feed preferences settings. Users can then choose to keep the current penalties, turn them off, or can even increase the level at which the platform penalizes such content.
RELATED LINKS:
Denuncian shadowban de grandes plataformas a contenidos sobre Palestina
Meta sancionado en Bélgica por aplicar “shadowban” a un eurodiputado
¿Fallas o censura oculta? polémica por restricciones a periodista palestino en X