Monthly Report January 2025

This month’s report highlights the key narratives driving hate speech across our baseline channels. By analysing online content across diverse digital platforms, including mainstream social media, alternative forums, and comment sections, we aim to provide actionable insights for combating online hatred and informing policy decisions.

Data for January was based on 2.95 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

The average toxicity in January 2025 was 0.21, showing a slight upward trend throughout the month.

Are you unsure about the terminology we use in this report? Visit our FAQ for more details.

Baseline channel analysis

The timeline graph illustrates the evolving toxicity levels concerning our baselines. The data reveals that toxicity toward Judaism remained consistently high at 0.34, while toxicity against Roma communities showed a slight increase from 0.27 to 0.31 in the first week of January. Islam-related toxicity fluctuated between 0.26 and 0.28, while LGBT+ toxicity experienced a decline from 0.32 to 0.27 before stabilising. Gender and refugee-related toxicity remained relatively lower, with minor variations.

Baseline Sexism

Volume
1,149,565 posts identified as toxic.

Toxicity
Average score was 0.18. Posts with extreme toxicity (0.8 or higher) totalled 13,131, predominantly on 4chan. 

Themes

• Sexism represents the most prominent hate category (28%), followed by Politics (15%), Violence (12%), Ridicule (11%), and Racism (11%).

• The 5 most common keywords in that channel were: Quran, cock, God, fuck and sex.

Baseline Anti-LGBTQ+

Volume
235,356 posts identified as toxic.

Toxicity
Average score was 0.29. Posts with extreme toxicity (0.8 or higher) totalled 1630, predominantly on X. 

Themes

• Politics (18%) and Obscenity (12%) represent the highest categories.

• Common toxic keywords: English terms (gay, women, sex), Spanish (gay, travesti), and Dutch (homo/gay)..

Key Insight
High concentration of gender-based hate speech and LGBTQ+ targeted content, with a significant toxicity score.

Baseline Anti-Muslim

Volume
326,434 posts identified as toxic.

Toxicity
Average score was 0.27. Posts with extreme toxicity (0.8 or higher) totaled 471, predominantly on X.

Themes

• Religion (68%) and Racism (43%) represent the highest categories

• Common toxic keywords:  English (quran, muslim, rape), Spanish (moro(s)/Moors), and Arabic terms: المشركين (polytheists), قتل (kill), الصهاينة (Zionists), كافر (infidel), الشيطان (Satan)

Key Insight
Very high volume of religious and racist violent content in Arabic. Large amount of messages were collected from X.

Baseline Anti-Refugee/Migrants

Volume
136,113 posts identified as toxic.

Toxicity
Average score was 0.23. Posts with extreme toxicity (0.8 or higher) totalled 1,211, predominantly on 4chan. 

Themes

• Politics (49%) and Racism (38%) represent the highest categories.

• Common toxic keywords: English (Trump, refugees, Israel), Spanish (Inmigrantes ilegales/Illegal immigrants),  Arabic (لاجئ/refugee)

Key Insight
Hate speech targeting migrants is flourishing in different languages.

Baseline Antisemitism

Volume
304,744 posts identified as toxic.

Toxicity
Average score was 0.34. Posts with extreme toxicity (0.8 or higher) totaled 5,698, predominantly on X. 

Themes

• Racism (75%), Religion (47%), and Politics (42%), represent the highest categories.

• Common toxic keywords: English (Jews), Arabic (حدود/borders), Spanish (judíos/Jews)

Key Insight
Hate speech targeting Muslims is flourishing in different languages.

Baseline Roma

Volume
36,621 posts identified as toxic.

Toxicity
Average score was 0.30. Posts with extreme toxicity (0.8 or higher) totalled 797, predominantly on X.

Themes

• Racism (52%) Politics (13%), and Religion (12%) represent the highest categories

• Common toxic keywords: Spanish (gitano(s)/gypsies), English (gypsies), Portuguese (os ciganos/the gypsies), Symbol (卐/swastika)

Key Insight
Anti-Semitic and anti-Roma content with Nazi symbolism suggests activity targeting multiple ethnic groups

This graph illustrates the distribution of toxicity levels—Neutral, Low, Medium, and High—across our monitored baselines. Sexism recorded the highest proportion of neutral messages, while antisemitic content had the largest share of medium toxicity messages (38.5%), closely followed by anti-Roma messages (38%). Highly toxic content, characterised by extreme hate speech (0.8–1 toxicity score), was most prevalent in messages targeting Jewish (2.8%), Roma (2.8%), and Muslim (2.6%) communities.

Hate speech by category

This table outlines key themes in hate speech across Gender, Islam, Judaism, LGBT+, Refugees, and Roma. Hate speech targeting refugees is the most politicised (49.4%), followed by Judaism (42.2%) and Islam (27.4%). Racism appears most frequently in hate speech about Judaism (75%) and Roma (52.3%), while the intersection of religion and Islam reaches 68.3%, highlighting intense religiously motivated hostility. The LGBTQ+ community faces the highest levels of sexism (58.9%), followed by gender (28.8%).