"Facebook Approved an Israeli Ad Calling for Assassination of Pro-Palestine Activist

After the ad was discovered, digital rights advocates ran an experiment testing the limits of Facebook’s machine-learning moderation. " - this was a test by NGOs

https://theintercept.com/2023/11/21/facebook-ad-israel-palestine-violence/

"A series of advertisements dehumanizing and calling for violence against Palestinians, intended to test Facebook’s content moderation standards, were all approved by the social network, according to materials shared with The Intercept.

The submitted ads, in both Hebrew and Arabic, included flagrant violations of policies for Facebook and its parent company Meta. Some contained violent content directly calling for the murder of Palestinian civilians, like ads demanding a “holocaust for the Palestinians” and to wipe out “Gazan women and children and the elderly.” Other posts, like those describing kids from Gaza as “future terrorists” and a reference to “Arab pigs,” contained dehumanizing language.

When compared to Arabic posts this is most likely no concidence.

"The approval of the Arabic versions of the ads come as a particular surprise following a recent report by the Wall Street Journal that Meta had lowered the level of certainty its algorithmic censorship system needed to remove Arabic posts — from 80 percent confidence that the post broke the rules, to just 25 percent. In other words, Meta was less sure that the Arabic posts it was suppressing or deleting actually contained policy violations.

Nashif said, “There have been sustained actions resulting in the silencing of Palestinian voices.”