Senator Elizabeth Warren (D-Mass) last week shared with Meta CEO Mark Zuckerberg her “concerns regarding recent reports that Meta is suppressing Palestinian and Palestinian-related content on Instagram.”
According to Warren, “Over 90 human rights and civil society organizations have raised alarm about Meta’s reported suppression, filtering, and mistranslation of Palestine-related content over the past two months, calling on the company to provide additional transparency as to content moderation and to address discriminatory algorithmic design.”
Despite our natural suspicion of the senator’s views and motives, this may be a good place to point out the harassment that has been endured over the years by right-wing Israeli users under Facebook’s biased policies.
In March 2022, then-only MK Itamar Ben-Gvir filed a lawsuit against Facebook alleging that the social network was “deliberately harming his exposure for political reasons.” According to Ben-Gvir, since he started his battle demanding safety for the Jewish residents of the Shimon Hatzadik (Sheikh Jarrah) neighborhood in Jerusalem, Facebook had reduced the exposure of his posts.
In June 2021, we reported that Facebook, Instagram, and Twitter, started to block the accounts and posts of right-wingers who support right-wing bloc parties, among them the prime minister’s son, Yair Netanyahu.
In 2017, the late blogger and IDF fighter Ari Fuld HY’D reported that his Facebook page had been suspended for a week over his comment: “All terrorists that have murdered Jews in Israel have been Muslims. Not all Muslims murder Jews but those that did were Muslim. I call them either radical Islamic terrorists or Arab terrorists.”
Back to Sen. Warren who continued: “Amidst the horrific Hamas terrorist attacks in Israel, a humanitarian catastrophe including the deaths of thousands of civilians in Gaza, and the killing of dozens of journalists, it is more important than ever that social media platforms do not censor truthful and legitimate content, particularly as people around the world turn to online communities to share and find information about developments in the region.”
Warren also complained that “news organizations have reported several instances of Instagram censoring or mistranslating Palestinian and Palestine-related content following Hamas’s deadly October 7 attacks.”
She noted that Instagram labeled the Palestinian flag emoji as “potentially offensive,” hiding from view comments containing said emoji. The platform also translated user bios that contained the Palestinian flag emoji, and the words “Palestinian” and “Hamdulilah” (Arabic for Praise be to God) as being “Palestinian terrorist,” with admittedly amusing outcomes such as, “Praise be to God, Palestinian terrorists are fighting for their freedom.”
“Meta reportedly manipulated its content filters to apply stricter standards to content generated in Palestine compared to the standards in place in the rest of the region and the world,” Warren claimed, adding: “Meta generally hides comments designated as hateful only when its content moderation systems are 80% certain that they violate the platform’s policies. The Wall Street Journal reported that after Hamas’s October 7 attacks, Meta tightened those filters, “as part of ‘temporary risk response measures.’”
Warren complained that “Meta cut the threshold in half over a swath of the Middle East, hiding any comment deemed 40% likely to be inflammatory. Meta also changed the default settings on Facebook to limit who could comment on content created ‘in the region.’ Soon thereafter, however, Meta lowered the bar further for Palestine only, hiding comments from users in the occupied Palestinian territories if its automated system judged there was at least a 25% chance they violated Meta’s rules.”
“Unsurprisingly, given these changes, Instagram users have reported numerous instances of account suspensions and content suppression linked to communications regarding Palestine,” she wrote.
According to Warren, while enhancing its censorship of “pro-Palestinian” messages, Facebook, which is owned by Meta, recently approved paid advertisements by a right-wing Israeli group “calling for the assassination of a pro-Palestine activist.” She also cited test ads that were submitted by 7amleh (pronounced Ḥamleh), calling for the “murder of Palestinian civilians,” which were also approved.
Sen. Warren concluded: “In order to further understand what legislative action might be necessary to address these issues, we ask for additional information about Meta’s application of its policies. Please answer the below questions no later than January 5, 2024 (she attached a 20-point questionnaire, including):
1. During the time period spanning October 7, 2023 to the present:
a. How many Arabic language posts originating from Palestine have been removed?
b. What percentage of total Arabic language posts originating from Palestine does the above number represent?
c. What percentage of the removed posts were removed due to automated systems versus human moderation?
d. How often did Meta limit the reachability of Arabic language posts originating from Palestine without notifying the user?
e. How often did Meta limit the reachability of Arabic language posts originating from Palestine while notifying the user?
2. During the time period spanning October 7, 2023 to the present:
a. How many English language posts originating from Palestine have been removed?
b. What percentage of total English language posts originating from Palestine does the above number represent?
c. What percentage of the removed posts were removed due to automated systems versus human moderation?
d. How often did Meta limit the reachability of English language posts originating from Palestine without notifying the user?
e. How often did Meta limit the reachability of English language posts originating from Palestine while notifying the user?
3. During the time period spanning October 7, 2023, to the present:
a. How many Hebrew language posts originating from Israel have been removed?
b. What percentage of total Hebrew language posts originating from Israel does the above number represent?
c. What percentage of the removed posts were removed due to automated systems versus human moderation?
d. How often did Meta limit the reachability of Hebrew language posts originating from Israel without notifying the user?
e. How often did Meta limit the reachability of Hebrew language posts originating from Israel while notifying the user?
4. During the time period spanning October 7, 2023, to the present:
a. How many English language posts originating from Israel have been removed?
b. What percentage of total English language posts originating from Israel does the above number represent?
c. What percentage of the removed posts were removed due to automated systems versus human moderation?
d. How often did Meta limit the reachability of English language posts originating from Israel without notifying the user?
e. How often did Meta limit the reachability of English language posts originating from Israel while notifying the user?
These questions continued over several pages, ending with:
20. What is Meta’s current policy for retention of content that may contain evidence of human rights abuses?
a. What are the criteria for initiating and terminating the preservation of content that may contain evidence of human rights abuses?
b. What is the period of data retention for content that may contain evidence of human rights abuses?
c. What are Meta’s policies for researcher, civil society, and governmental access to content that may contain evidence of human rights abuses?
Come January 5, we should find out just how much worse Facebook and the rest of Meta will have become after having to comply with Sen. Warren’s inquiry. You know they won’t get better.
