CyberWell Study Reveals Disparity inHolocaust Hate SpeechRemoval on Social Media

A recent study by CyberWell has uncovered a significant disparity in the removal of hate speech posts on social media platforms. The study, which analyzed posts from April 2023 to April 2024 across Facebook, Instagram, TikTok, YouTube, and X, found that only 9% of Arabic posts with Holocaust-related hate speech were removed, compared to 35% of similar English posts.

Why this matters: This disparity in hate speech removal highlights the need for social media companies to re-examine their content moderation practices and ensure consistency across all languages and regions. Failure to address this issue can perpetuate harmful ideologies and contribute to a rise in online hate crimes.

The stark difference raises concerns about the inconsistent application of hate speech policies across languages and regions. The study's findings suggest that social media platforms may be more lenient in removing hate speech content in Arabic, potentially contributing to the perpetuation of harmful and offensive ideologies.

The removal rate for Arabic posts was lower across all five platforms examined. On Facebook, 11% of English posts denying or distorting the Holocaust were removed in 2023, and 17% in 2024, while only 4% of similar Arabic posts were removed in 2023, and less than 1% in 2024. On X, 34% of English Holocaust distortion posts were removed in 2023, and 42% in 2024, while only 8% of similar Arabic posts were removed in 2023, and 3% in 2024.

The majority of posts identified (61%) breached the International Holocaust Remembrance Alliance (IHRA) definition of antisemitism under article 4, Holocaust denial. 26% of posts accused Jews as a people or Israel as a state of inventing or exaggerating the Holocaust. 30% of posts included stereotypes, generalizations, and conspiratorial content about Jews. 24% of posts called for glorified or justified violence against Jews, and 8% of posts compared Israel to the Nazis.

One post used the term "Holohoax" and claimed Jews were to blame for the Russian Revolution and the deaths of 30 million Russians. The post gained over 16,000 views despite being flagged on X with the tag "Visibility Limited". The phrase "Hitler was right" was used over 2,600 times, with a potential reach of 1.8 million users since the beginning of April 2024.

The study noted that during times of heightened conflict between Israel and Hamas, posts comparing Jews with Nazis increase, and the phrase "Hitler was right" is commonly used. CyberWell also detected a significant number of antisemitic posts that use "Zionist", "Zionists", or "Zios" as a coded term for Jews.

The CyberWell study highlights the need for social media companies to re-examine their content moderation practices and ensure that they are consistently enforcing their hate speech policies across all languages and regions. The findings underscore the importance of combating hate speech and promoting online safety, particularly when it comes to non-English languages.

Key Takeaways

• Only 9% of Arabic posts with Holocaust-related hate speech were removed, vs 35% of similar English posts.
• Removal rate for Arabic posts was lower across all 5 social media platforms examined.
• 61% of identified posts breached the IHRA definition of antisemitism, including Holocaust denial and stereotypes.
• Antisemitic posts increased during times of conflict between Israel and Hamas, using coded terms like "Zionist" for Jews.
• Social media companies must re-examine content moderation practices to ensure consistency across languages and regions.