7amleh and the international civil-society campaign on Meta's content moderation of Palestinian content (2023–ongoing)

The campaign anchored by 7amleh — the Arab Center for the Advancement of Social Media, the Haifa-headquartered Palestinian digital-rights organisation — and built up with Human Rights Watch, Access Now, Amnesty International, Article 19, the Electronic Frontier Foundation, Witness, and the wider Palestinian and international civil-society field is the corpus's principal grassroots-organising response to AI-mediated platform content moderation under wartime conditions. Its substantive register is the asymmetric pattern by which Meta's automated content-moderation systems on Facebook, Instagram, and WhatsApp have, since the 7 October 2023 escalation of the Israel-Gaza war, over-removed and over-restricted Arabic-language and pro-Palestine content while under-enforcing Hebrew-language incitement and violent content against Palestinians. It is the corpus's first Palestinian civil-society campaign, the first WANA / MENA-anchored campaign of any kind, and — distinct from the Kenya-anchored AI-supply-chain labour cases of Foxglove and the 185 former Facebook content moderators and the Daniel Motaung petition — the corpus's first instance of a campaign engaging the affected user-side of the AI-mediated content-moderation pipeline rather than its outsourced-labour side.

Origins: the May 2021 escalation and the BSR Human Rights Impact Assessment

The campaign's documentary roots predate the post-October 2023 intensification. The May 2021 Israel-Gaza escalation surfaced widespread takedowns of Palestinian and Arabic-language content on Facebook and Instagram, and the Meta Oversight Board recommended an independent human-rights impact assessment of Meta's policies during the escalation. Meta commissioned Business for Social Responsibility (BSR), whose September 2022 report produced 21 specific recommendations, explicitly found over-enforcement of Arabic-language content compared with Hebrew, and cited adverse human-rights implications for Palestinians' freedom of expression, freedom of assembly, political participation, and non-discrimination. 7amleh's 27 September 2022 statement on the BSR review — jointly signed with Human Rights Watch, Article 19, and regional partners — pressed Meta for full implementation of the BSR recommendations and is the campaign's pre-October 2023 multi-organisation anchor. The substantive shift after October 2023 is therefore less the documentary opening of the file than the move of an already-existing documentary file into a wartime register that, on the documentary record built up since, has demonstrated both the scale of the under-enforced incitement against Palestinians and the corresponding scale of the over-enforced moderation of Palestinian voices.

Post-October 2023: the Oversight Board, Human Rights Watch, and Access Now

The campaign's first watershed moments after October 2023 are two near-simultaneous external validations of 7amleh's pre-existing documentary case. On 19 December 2023 the Meta Oversight Board issued its expedited decision in the Al-Shifa Hospital case (Case IG-WUC3649N), overturning Meta's automated removal of a Palestinian-civilian-casualty Instagram post about the strike on or near Al-Shifa Hospital in Gaza. The Board found that both the initial removal and the rejection of the user's appeal had been taken automatically by classifier without any human review, identified Meta's lowering of its Violent and Graphic Content removal threshold during the post-October 7 crisis response as a contributing cause, and expressed concern about unintentional bias in Meta's practices against Palestinian and Arabic-speaking users — the first formal Oversight-Board acknowledgement of the over-enforcement pattern that anchors the 7amleh campaign's documentary case. Two days later, on 21 December 2023, Human Rights Watch released "Meta's Broken Promises: Systemic Censorship of Palestine Content on Instagram and Facebook", the international-human-rights field's most substantial post-October 2023 audit of Meta's content moderation of Palestinian content. HRW verified 1,050 cases of pro-Palestine content-moderation actions reviewed between October and November 2023, identified six recurring censorship patterns (post removals, account suspensions, engagement restrictions, follow/tag restrictions, feature restrictions, shadow-banning) each appearing at least 100 times, named 7amleh as a civil-society partner alongside Access Now and Amnesty International, and put six headline recommendations to Meta on the public record — overhaul of the Dangerous Organizations and Individuals (DOI) policy, transparency on government takedown requests, audit of the "newsworthiness" allowance for equitable application, disclosure of automation and algorithm error rates, ensuring functional appeal mechanisms for all users globally, and human-rights due-diligence on temporary algorithmic changes. The HRW report has been carried in international and regional press from Middle East Monitor's 21 December 2023 coverage outward as the international civil-society field's authoritative statement of the documentary case.

Access Now's 19 February 2024 report "It's not a glitch: how Meta systematically censors Palestinian voices", by MENA Policy and Advocacy Director Marwa Fatafta, deepened the audit. The Access Now report documented Meta's lowering of its algorithmic detection thresholds during the post-October 2023 crisis-response window — from 80% to 40% for Middle East content, and to 25% specifically for Palestine — surfaced Meta's leaked 2020 internal documents showing a 77% error rate for Arabic-language terrorist-content classifiers on non-violent content, and documented Meta's advertising systems approving paid advertisements calling for assassination and genocide against Palestinians. Access Now's data base for the report drew principally on 7amleh's 7or — the Palestinian Observatory of Digital Rights Violations — which had documented approximately 1,043 instances of censorship between 7 October 2023 and 9 February 2024. The HRW and Access Now reports together — with 7amleh's monitoring infrastructure underneath both — comprise the campaign's external-validation backbone in the first six months of the war.

The 7amleh documentary record: Hashtag Palestine, testimony, and the cross-platform expansion

7amleh's own annual and topical reporting carries the campaign's primary documentary load. The 9th annual Hashtag Palestine report, released 17 January 2024, documented 4,400 digital-rights violations through the 2023 calendar year, finding that 69% of them occurred after 7 October 2023, and attributing the surge to Meta's automatic-moderation systems discriminating against Arabic-language content. The 16 September 2024 "Palestinian Digital Rights, Genocide, and Big Tech Accountability" report — released to mark one year of the war — documented over 5,100 digital-censorship cases across Meta and X between 7 October 2023 and September 2024, documented Meta's targeted advertising approving content calling for assassination and forced Palestinian expulsion, and broadened the campaign documentary record from Meta to a cross-platform register including YouTube's hosting of Israeli Ministry of Foreign Affairs war-promotion advertising (a $7.1 million budget in two weeks post-October 2023 targeting France, Germany, and UK audiences). The 18 December 2024 "Erased and Suppressed: Palestinian Testimonies of Meta's Censorship" compiled 20 first-person testimonies from Palestinian influencers, journalists, and media outlets documenting post deletions, account suspensions, and reach restrictions on Facebook and Instagram, framing the harms in economic, professional, psychological, social, and political terms — the campaign's most direct entry of affected-user voices into the international documentary record. The Hashtag Palestine 2024 report, "The War on Gaza, Digital Rights Violations, and Weaponization of AI", released 27 January 2025, placed the platform-content-moderation record alongside documentation of approximately 75% destruction of Gaza's internet and communication infrastructure and a parallel record of Israeli AI-targeting systems Lavender and Habsora, sustaining the campaign's broader AI-and-armed-conflict register through the 15-month-plus war. The 2 September 2025 "Meta's Role in Amplifying Harmful Content During Genocide" report — covered internationally in IFEX and other free-expression press — documented more than 2,000 pieces of harmful Hebrew-language content across Facebook, Instagram, and WhatsApp between October 2023 and May 2025, detected 2.5 million harmful Hebrew posts in 2024 alone, characterised Meta's asymmetric enforcement as "systematic racial discrimination in content governance", and documented Meta's advertising systems approving violent Hebrew-language content as paid advertising in direct experiments in 2023 and 2025.

The Violence Indicator: civil-society AI tooling at the scale of the platform

7amleh's in-house AI-monitoring infrastructure — branded the Violence Indicator — is the campaign's clearest operational artefact. The tool is an AI-powered language-model analysis system that has detected more than 15 million inciting Hebrew-language posts against Palestinians on social media since October 2023, sized at approximately 3 million instances in the Hashtag Palestine 2023 baseline and 2.5 million across the 2024 calendar year alone in the September 2025 report. The Violence Indicator is the corpus's clearest documented case of grassroots Arabic-language civil-society building its own AI tooling at the scale of the platform whose moderation practices it is contesting — an operational inversion in which the civil-society field assembles the same kind of large-scale automated content-classification capacity Meta itself uses, but turns the classifier outward to detect what Meta's own systems are not removing. Together with 7or, the Palestinian Observatory of Digital Rights Violations that 7amleh runs as the campaign's case-by-case documentation system, the Violence Indicator gives the campaign a documentary record at a scale and level of detail that has anchored every external civil-society and international-press validation of the asymmetric-moderation pattern since.

Coalition register: international civil society, multistakeholder venues, and the Palestinian field

The campaign's coalition register operates at three concentric levels. Inside Palestinian civil society, 7amleh anchors the Palestinian Digital Rights Coalition and convenes the Palestine Digital Activism Forum, the annual Arabic-language and Palestinian-civil-society convening on digital rights — the 8th edition of which was held 4–5 June 2024 in partnership with SMEX's Bread & Net 6th edition under the theme of artificial intelligence's role in human rights, pairing Palestinian civil society with the wider WANA regional Arabic-language digital-rights field. Inside the international civil-society field, 7amleh works alongside Human Rights Watch, Access Now, Amnesty International, Article 19, the Electronic Frontier Foundation, and Witness on joint statements, coordinated reporting, and shared engagement with Meta and the Oversight Board. Inside the multistakeholder platform-accountability infrastructure, 7amleh is a civil-society member of the Global Network Initiative, holds UN ECOSOC consultative status, and engages directly with the Meta Oversight Board on Palestinian-content cases — a posture distinct from the corpus's other content-moderation campaigns, whose register is litigative rather than multistakeholder. The coalition's framing across these three levels is that platform-content moderation of Palestinian content during the Israel-Gaza war is not a glitch, an inadvertent error, or an unintended consequence of crisis response, but a systemic and documentable pattern of asymmetric enforcement that admits of policy reform and audit-engagement remedies — the substantive theory of change that distinguishes 7amleh's campaign from a litigation-side response.

Significance for the broader AI-good movement

The campaign is the corpus's most substantial mapped instance of grassroots-civil-society engagement with AI-mediated content moderation on the affected-user side of the pipeline. Where the Kenya-anchored 185-moderators and Motaung petitions test the labour-side of Meta's outsourced moderation chain, the 7amleh campaign tests the platform-end-product side of the same AI-content-moderation infrastructure: the wartime application of automated removal classifiers, the lowering of crisis-response thresholds, the asymmetric outcomes for Arabic-language and Hebrew-language content, and the inequity of access to functional appeal mechanisms. It is the corpus's first campaign to engage Palestinian users, journalists, and media outlets as non-AI publics whose engagement is itself the substantive participatory register the mission tracks — affected populations contesting how AI is being deployed against them, building their own AI-monitoring infrastructure in response, and translating the documentary record into international audit-engagement and human-rights-mechanism work. The campaign also sits in a load-bearing relationship to the wider AI-and-armed-conflict register: its post-October 2023 file is the corpus's clearest demonstration that wartime conditions do not depopulate the grassroots make-AI-good field, but reshape its substantive register — from peacetime platform-policy engagement to wartime documentation of AI-mediated harm at scale, with the same civil-society infrastructure and the same multistakeholder venues carrying the work through.