A new audit published by the World Jewish Congress (WJC) claims that the vast majority of antisemitic content reported by European Jewish community members remains publicly available online, pointing to what the organization characterizes as profound gaps in platform moderation and accountability.
The study, titled “Flagged and Failed: Platform Responses to Antisemitic Content Reported by Jewish Communities,” evaluates data collected between February 2025 and April 2026. The initiative was carried out under the WJC Hate Online: Preparedness and Empowerment (HO:PE) project, which was co-funded by the European Union under the CERV program.
According to the organization, this marked the first systematic, community-sourced dataset of online antisemitism gathered across the EU via a purpose-built reporting application.
The project used a dedicated reporting tool that enabled Jewish community members across Europe to document antisemitic URLs, categorize them using the International Holocaust Remembrance Alliance’s working definition of antisemitism as an analytical framework, and indicate whether they had formally flagged the posts to host platforms.
Following the filtering of invalid entries, the published dataset comprised 754 distinct reports spanning six major social media platforms. To measure platform accountability, researchers conducted a liveness check throughout May 2026 to assess whether the reported content was still publicly accessible.
Furthermore, a specialized legal assessment was conducted for high-risk categories to evaluate potential thresholds for illegality under EU and member state laws, including the EU Framework Decision on Racism and Xenophobia.
Online hate spans a wide variety of tropes
The analysis presented by the organization demonstrates that online hate spans traditional tropes, Holocaust distortion, incitement to violence, and hostility related to Israel.
Facebook accounted for the highest volume of submissions at 34.4%, followed closely by Instagram and X/Twitter at 26% each, and TikTok at 9.7%. When examining how content remains online, the audit finds striking disparities in how platforms handle reported material.
Across all submissions evaluated in May 2026, the study contends that 81.3% of reported antisemitic posts, equaling 613 out of 754, remained live. A critical assertion of the study is that formally reporting content through a platform’s built-in mechanism made virtually no measurable difference in whether it was removed.
The overall removal rate hovered near 19% for content that was formally flagged, compared to roughly 18% for content left unflagged.
Platform-specific breakdowns highlight diverse enforcement realities: Facebook recorded the highest user-reporting rate in the dataset, with 76% of its 259 entries formally flagged by users, yet registered the lowest overall removal rate among major platforms at just 13.9%.
Instagram and X both showed removal rates of 12%-15%, with users reporting a relatively small percentage of encountered content directly to the platforms.
Standing apart as an outlier, TikTok removed 64.4% of all content in its dataset and 70.6% of content explicitly reported to it, making it the only major platform where the study recorded a majority of audited antisemitic posts being taken down.
The report’s category analysis illustrates that certain types of hate speech are vastly more resistant to moderation than others. Israel-related antisemitism proved to be the most persistent blind spot for moderators according to the data, with only 5.2% of such reports removed.
Content comparing Israeli policy to Nazism showed no removal across all reports
Most notably, the audit claims that content comparing Israeli policy to Nazism showed no removal across all 28 documented reports.
Conversely, explicit incitement to violence, such as calls to harm Jewish people or radical ideological incitement, carried the highest removal rate at nearly 35%, though dangerous material like terrorist propaganda and direct calls to murder still frequently lingered online. Overall, the legal assessment estimated that roughly 6% of all submissions crossed thresholds warranting individual criminal examination under European law.
The WJC concluded that self-regulatory reporting mechanisms are failing everyday users, leaving vulnerable communities exposed to persistent digital hate. In response, the report outlines direct recommendations for social media companies to audit their content moderation policies, demonstrate the operational efficacy of their reporting channels with transparent data, and address blind spots regarding conflict-related hate speech.
Regulatory bodies are also urged to use enforcement powers under instruments such as the Digital Services Act to scrutinize risk-mitigation practices. In contrast, national law enforcement agencies must establish clear referral pathways for illegal online incitement.