Why Social Media Sites Failed to Evade Censorship and Suppress Hate Speech and Disinformation During the Gaza War
LONDON: Tech giant Meta has announced it will begin removing social media posts that use the term “Zionist” in the context of Jews and Israelis, as part of an effort to curb anti-Semitism on its platform.
The parent company of Facebook and Instagram previously announced it would lift a blanket ban on the most-moderated term across all of its platforms on Meta, “shaheed,” or “martyr,” after a year-long review by its oversight board concluded the approach was “too broad.”
Likewise, TikTok, X, and Telegram have long pledged to step up efforts to curb the spread of hate speech and misinformation on their platforms amid the ongoing war in Gaza.
These initiatives aim to create a safer, less toxic online environment. But as experts have repeatedly pointed out, these efforts often fall short, leading to empty promises and worrying trends in censorship.
“Simply put, social media platforms have not been very good at avoiding censorship of the Gaza war or suppressing hate speech and disinformation,” Nadim Nassif, founder and director of 7amleh, a digital rights and human rights activist group for Palestinians, told Arab News.
“Censorship and account deletions throughout the conflict have also jeopardized efforts to document human rights abuses on the ground.”
Nassif said hate speech and incitement to violence remain “rampant,” particularly on Meta and X, where anti-Semitic and Islamophobic content “continues to be widespread.”
Since the Hamas-led Gaza conflict began on October 7, social media has been awash with war-related content. In many cases, social media has served as a vital window into the dramatic events unfolding in the region, and has become a vital source of real-time news and accountability for Israel’s actions.
Profiles supporting both Hamas and the Israeli government's actions have been criticized for sharing misleading and hateful content.
fastactually
1,050
Instances of Palestinian and supporter posts being deleted or otherwise suppressed on Instagram and Facebook, documented by Human Rights Watch between October and November 2023.
Yet no social media platform, including Meta, YouTube, X, or messaging apps like TikTok or Telegram, has publicly outlined policies designed to moderate hate speech and violence related to the conflict.
Instead, these platforms continue to be flooded with war propaganda, dehumanizing rhetoric, genocidal rhetoric, explicit calls to violence, and racist hate speech. In some cases, the platforms have removed pro-Palestinian content, blocked accounts, and sometimes shadowbanned users who express support for the Gazan people.
On Friday, Turkey’s telecommunications authority blocked access to Instagram, a social media platform owned by Meta. Local media said the blocking was in response to Instagram deleting posts by Turkish users expressing condolences over the recent killing of Hamas political leader Ismail Haniyeh in Tehran.
The previous day, Malaysian Prime Minister Anwar Ibrahim had denounced Mehta as cowardly after his Facebook post about Haniyeh’s killing was deleted. “Let this be a clear and unambiguous message to Mehta: Stop your cowardly actions,” Anwar, who has repeatedly criticized Israel’s war in Gaza and its actions in the occupied West Bank, wrote on his Facebook page.
Videos of Israeli soldiers blowing up mosques and houses, burning Qurans, torturing and humiliating blindfolded Palestinian prisoners, dragging them around on the hoods of military vehicles, and glorifying war crimes are still available for free on mobile screens.
“Historically, the platform has been very bad at moderating content about Israel and Palestine,” Nassif said. “With the Gaza war and the plausible genocide that is currently underway, that has simply gotten worse.”
A Human Rights Watch report published in December, “Meta’s Broken Promises,” accused the company of “systematic online censorship,” “inconsistent and opaque application of its policies,” and silencing voices supporting Palestine and Palestinian rights on Instagram and Facebook.
The report added that Meta's actions “failed to meet its human rights due diligence obligations” due to years of failure to address “overly broad enforcement”.
“I’m not sure to what extent this can really be called an effort to end censorship,” Jacob Mukherjee, convener of the Masters in Political Communication at Goldsmiths, University of London, told Arab News.
“Meta has promised to conduct various reviews. By the way, this is something that has been promised for more than two years, since the last escalation of the Israeli-Palestinian conflict in 2021. Before October 7th of last year.
“But I don't think it's really changed much, really. They've had to respond to suggestions that they were engaging in censorship, but I think it's mostly been a public relations exercise.”
Between October and November 2023, Human Rights Watch documented more than 1,050 instances of the removal and suppression of posts by Palestinians and their supporters on Instagram and Facebook, including content about human rights abuses.
Of these, 1,049 were censored or unfairly suppressed for peaceful content supporting Palestine, and one was for the removal of content supporting Israel.
But censorship appears to be only part of the problem.
7amleh's Violence Barometer, which monitors real-time data on violent content in Hebrew and Arabic across social media platforms, has recorded more than 8.6 million violent content items since the conflict began.
Nassif says the surge in violent and harmful content, primarily in Hebrew, is largely due to a lack of investment in moderation.
This content, which primarily targets Palestinians on platforms such as Facebook and Instagram, was used as evidence by South Africa in its case against Israel at the International Court of Justice.
It is not just Mehta who is responsible for what South African lawyers have described as the first mass killing to be broadcast live on mobile, computer and TV screens.
X has also faced criticism for its liberal use of hashtags known to spread misinformation and manipulated images, both from Palestinian and Israeli supporters, a practice often shared by high-profile politicians and media figures.
“One of the main problems with the current content moderation system is the lack of transparency,” Nassif said.
“When it comes to AI, platforms don’t disclose clear and transparent information about how and when AI systems are implemented in their content moderation processes. Policies are often opaque, leaving platforms with a lot of leeway to do as they see fit.”
According to Mukherjee, the issue of moderation behind the smokescreen of ambiguous policies is a very political one, and these companies need to adopt a “balanced” approach between political pressure and “managing the expectations and desires of their user base.”
“These AI tools could be used to shield the people who run the platforms, the real power holders, from criticism and accountability, and that’s a real problem,” he said.
“These platforms are essentially private monopolies responsible for regulating important parts of the political public sphere.
“In other words, they help shape and regulate the arena in which conversations take place, the arena in which people form opinions, the arena in which politicians feel the pressure of public opinion, but they are not accountable at all.”
As Arab News reported in October, there have been instances of pro-Palestinian content being censored or removed, but these platforms made it clear long before the Gaza conflict that removing content from their platforms was ultimately not in their interest.
“These platforms are not built for the public good, nor are they meant to create an educated public that is exposed to diverse viewpoints and has the information to make informed decisions and form opinions,” Mukherjee said.
“The reality is (the business model) is you want to have a lot of content, and if it's pro-Palestinian content, then do that. Ultimately, it's still about getting eyeballs on the platform and driving engagement, and in industry jargon, content that evokes strong emotions is what drives engagement, and that means data, and that means money.”