Monthly Report April 2025

Data for April was based on 2,797,266 M messages in 11 languages, across 6 social media platforms: Reddit, X, 4chan, Gab, YouTube, and Facebook.

Content warning: Presented data may contain disturbing language related to online hate speech.

Average toxicity

In April, the overall average toxicity level was 0.20. This figure saw a slight uptick toward the end of the month, reaching 0.22 before dropping to 0.19 on the final day.

Baseline channel analysis

This section provides an overview of the average toxicity levels among the baselines over April. Antisemitism remains the category with the highest average, maintaining a stable score of 0.33 throughout the month. Toxicity in messages targeting the Roma community showed significant fluctuations, peaking at an average score of 0.35 both the first and last weeks of April. Anti-Muslim and anti-LGBTQ+ narratives maintained a relatively stable trend, both with an average of 0.27. Sexist content had the lowest toxicity, while anti-refugee narratives experienced a modest increase in the latter half of the month.

Baseline Sexism

Volume: 307,976 posts identified as toxic content.

Platforms: Most posts (75%) originated from X, with Reddit (10%) and TikTok (9%) also contributing significantly.

Toxicity: The average toxicity score was 0.18. Posts with extreme toxicity (0.8 or higher) totaled 17,748, predominantly on 4chan.

Themes:
Sexism was the most prevalent category (46%), followed by Violence (31%).

Common toxic keywords included bitch (English), slut (English), حد (Arabic; "limit" or potentially a reference in religious/legal context), whore (English), and freak (English).

Key Insight: Toxicity is concentrated on 4chan and disproportionately targets women, with nearly half of the most harmful content rooted in sexist language.

Baseline Anti-LGBTQ+

Volume: 94,108 posts identified as toxic content.

Platforms: Most posts (84%) originated from X, with Reddit (9%) and TikTok (3%) also contributing significantly.

Toxicity: The average toxicity score was 0.29. Posts with extreme toxicity (0.8 or higher) totaled 4,435, predominantly on Gab.

Themes:
Sexism was the most prevalent category (75%), followed by Violence (20%).

Common toxic keywords included شاذ (Arabic; “pervert” or “deviant”), slut (English), bitch (English), pqp (Portuguese abbreviation for a strong insult), and puto (Spanish; offensive slang, often homophobic).

Key Insight: Although the majority of content comes from X, Gab hosts the highest concentration of extremely toxic posts, with misogynistic and homophobic slurs dominating the discourse.

Baseline Anti-Muslim

Volume: 129,804 posts identified as toxic content.

Platforms: Most posts (89%) originated from X, with TikTok (4%) and Reddit (3%) also contributing significantly.

Toxicity: The average toxicity score was 0.27. Posts with extreme toxicity (0.8 or higher) totaled 10,916, predominantly on 4chan.

Themes:
Religion was the most prevalent category (76%), followed by Violence (36%).

Common toxic keywords included moros (Spanish; derogatory term for Muslims), moro (Filipino/Spanish; pejorative for Muslim groups), المشركين (Arabic; “polytheists,” often used in a hostile religious context), mora (Italian/Spanish slang; can carry xenophobic connotations), and الصهاينة (Arabic; “Zionists”).

Key Insight: Religious hate speech is the dominant toxic theme, with 4chan showing the highest concentration of extreme content targeting Muslims and perceived religious enemies.

Baseline Anti-Refugee/Migrants

Volume: 43,929 posts identified as toxic content.

Platforms: Most posts originated from X (88%), with Reddit (4%) and Facebook (4%) also contributing significantly.

Toxicity: The average toxicity score was 0.22. Posts with extreme toxicity (0.8 or higher) totaled 1,497, predominantly on 4chan.

Themes:
Politics was the most prevalent category (65%), followed by Violence (30%).

Common toxic keywords included fachkräfte (German; sarcastic use meaning “skilled workers,” often referring to migrants), "migrants" (English; in scare quotes to signal hostility), immigrati clandestini (Italian; “illegal immigrants”), negro (Spanish/Italian; racial slur), and vergewaltiger (German; “rapist”).

Key Insight: Political hate speech, especially targeting migrants, is dominant, with extreme toxicity concentrated on 4chan and marked by xenophobic and racist terminology.

Baseline Antisemitism

Volume: 156,079 posts identified as toxic content.

Platforms: Most posts (89%) originated from X, with Reddit (3%) and 4chan (3%) also contributing significantly.

Toxicity: The average toxicity score was 0.34. Posts with extreme toxicity (0.8 or higher) totaled 10,043, predominantly on 4chan.

Themes:
Racism was the most prevalent category (75%), followed by Violence (41%).

Common toxic keywords included الصهاينة (Arabic; “Zionists”), kill jews (English), nazista (Portuguese; “Nazi”), قتل (Arabic; “kill”), and niggers (English slur).

Key Insight: Despite X being the primary source of posts, the most toxic content clusters on 4chan, where highly racist and violent rhetoric dominates.

Baseline Roma

Volume: 16,913 posts identified as toxic content.

Platforms: Most posts originated from X (83%), with Reddit (6%) and YouTube (4%) also contributing significantly.

Toxicity: The average toxicity score was 0.27. Posts with extreme toxicity (0.8 or higher) totaled 864, predominantly on Gab.

Themes:
Racism was the most prevalent category (84%), followed by Violence (15%).

Common toxic keywords included gitanos (Spanish; “gypsies,” often used pejoratively), gitano (Spanish; singular form), zigeuner (German; derogatory term for Roma people), moros (Spanish; slur for Muslims), and os ciganos (Portuguese; “the gypsies,” often used negatively).

Key Insight: Racist hate speech targeting Roma and Muslim communities is dominant, with Gab hosting the highest share of extremely toxic content.

This graph illustrates the distribution of toxicity levels—Neutral, Low, Medium, and High—across our monitored baselines. Anti-Roma and Antisemitism narratives recorded the highest shares of medium and high toxicity, with 38.56% and 37.27% at the medium level, and 3.27% and 2.44% at the high level, respectively. Anti-Muslim content also showed elevated toxicity, with 2.54% of messages rated highly toxic. In contrast, sexism-related content was the least toxic overall, with over 60% of the messages rated as neutral, followed by anti-refugee narratives at 45.48%.

Hate speech by category

The table highlights the thematic intersections most frequently associated with each monitored baseline. Racist language was most prevalent in content targeting the Jewish community (73.19%), followed by the Roma (54.4%) and Muslim (42.13%) communities. Political discourse showed a strong overlap with anti-refugee (46.47%) and antisemitic (41.14%) narratives. Religion-related hate was most prominently directed at Muslims (67.8%), with Jewish people also significantly targeted (51.46%). Meanwhile, sexist language was most commonly directed at the LGBTQ+ community, appearing in 57.17% of related messages.