Meta's AI moderation systematically suppressed Palestinian voices, study finds
A 2021-2025 study by 7amleh and Utrecht University documented 3,520 cases of account suspensions, content removals, and reduced visibility affecting Palestinian users and journalists on Meta platforms. The report argues that repeated restrictions on Palestinian content cannot be explained as isolated errors, though it does not conclude that Meta deliberately censored individuals.
What happened
The Platformicide of Palestine 2021β2025 report, published by 7amleh and a researcher at Utrecht University, examined 3,520 cases submitted to 7amleh's Palestinian Observatory for Digital Rights Violations from January 2021 to December 2025.
The documented cases included account suspensions, post removals, restrictions on livestreaming and reduced content distribution. Only 32.5% of the cases involved an identifiable policy violation.
Among cases where a reason was provided, Meta's Dangerous Organizations and Individuals policy accounted for 57.2% of restrictions, rising to 71.1% in cases involving journalists. The policy is intended to prevent support for terrorist and violent organizations, but the report argues it has been applied too broadly to journalists, media organizations, activists and civil society groups covering Palestine and Israel's war on Gaza.
Reported impact
Jerusalem-based journalist Faten Elwan, who has more than 300,000 followers on Instagram, saw her posts drop from 30,000 views to a few hundred within an hour after her account was reactivated in April 2026 following a two-month suspension. Elwan's account was suspended eight times during the war and she was unable to find assistance when she met with Meta representatives in May 2025.
The report uses the term 'platformicide' to describe the structural erasure of Palestinian visibility and participation on social media. Account suspensions were the most common enforcement measure, affecting about one-third of the documented cases, while less visible restrictions included reduced reach and recommendation filtering, commonly described as 'shadow banning'.
Index analysis
Nearly half of the 2,828 appeals submitted by 7amleh on behalf of affected users received no response from Meta. The report argues that the repeated restrictions cannot be explained solely as isolated technical or moderation errors.
The report calls on Meta to commission an independent human rights audit, reform its application of the Dangerous Organizations and Individuals policy and strengthen protections for journalistic and public-interest content.
Governance controls
Meta has previously said it was working to improve the accuracy of Arabic-language moderation, route content to reviewers familiar with relevant Arabic dialects and provide users with clearer explanations of enforcement decisions.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.