The 69th Eurovision Song Contest, held in Basel, Switzerland, from May 13 to 17, 2025, turned out to be a politically charged edition. Austria's JJ clinched victory with the operatic pop ballad 'Wasted Love,' but the spotlight was equally on Israel's Yuval Raphael, who won the public vote amid allegations of vote manipulation and geopolitical tension.

Data for this report was based on 31,094 posts in English, German, French, Spanish, Dutch, and English across 8 social media platforms: Reddit, X, 4chan, Gab, YouTube, TikTok, 9Gag and Facebook.

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

Raphael, a survivor of the 2023 Nova music festival attack, performed 'New Day Will Rise' and finished second overall. Her participation sparked protests, accusations of government-led vote campaigns, and criticism of the European Broadcasting Union (EBU) for allegedly masking audience boos with fake applause during rehearsals.

The contest also faced internal dissent. Malta was pressured to alter their song lyrics ahead of the contest after a complaint by the BBC, raising concerns about censorship. Meanwhile, Estonia's entry, 'Espresso Macchiato' by Tommy Cash, drew ire in Italy for its satirical portrayal of Italian stereotypes, leading to calls for disqualification. These controversies show the increasingly politicised nature of Eurovision, both on stage and online.

Multilingual toxicity analysis

To understand the digital discourse surrounding Eurovision 2025, we analysed posts containing the keywords 'Eurovision' and 'ESC' across six languages: English, German, French, Spanish, Dutch, and Italian. We selected these six languages because they represent the largest and most robust datasets in our system, allowing us to compare how online discourse around Eurovision differs across linguistic and cultural contexts. The data was collected between May 8 and May 20, 2025.

English

This channel contains 12,473 posts, of which 1,144 are toxic.

Posts were collected between May 08, 2025 and May 20, 2025.
Most data was collected from Twitter (45%), Tiktok (42%) and/or Reddit (11%).
Posts are all written in English (100%).
The average toxicity score is 0.10.
10 posts have a toxicity score of 0.8 or higher.
Messages with high toxicity are more frequent on 4chan.
Toxicity often involves Politics (60%).
About 11% of messages involve Violence.
Keywords frequently found in toxic messages include: bitch, freaks, freak, whore, and morons.

German

This channel contains 4,212 posts, of which 652 are toxic.

Posts were collected between May 08, 2025 and May 20, 2025.
Most data was collected from Twitter (91%), Tiktok (5%) and/or Youtube (2%).
Posts are all written in German (100%).
The average toxicity score is 0.13.
7 posts have a toxicity score of 0.8 or higher.
Messages with high toxicity are more frequent on YouTube.
Toxicity often involves Politics (37%).
About 6% of messages involve Violence.
Keywords frequently found in toxic messages include: dreck, verpiss dich, narr, fette, and pack.

French

This channel contains 4,314 posts, of which 823 are toxic.

Posts were collected between May 08, 2025 and May 20, 2025.
Most data was collected from Twitter (96%), Tiktok (3%) and/or Reddit (1%).
Posts are all written in French (100%).
The average toxicity score is 0.19.
18 posts have a toxicity score of 0.8 or higher.
Messages with high toxicity are more frequent on Twitter.
Toxicity often involves Racism (64%).
About 4% of messages involve Violence.
Keywords frequently found in toxic messages include: israhell, foutre, dehors, fout, and branle.

Spanish

This channel contains 4,700 posts, of which 647 are toxic.

Posts were collected between May 08, 2025 and May 20, 2025.
Most data was collected from Twitter (92%), Tiktok (7%) and/or Youtube (1%).
Posts are all written in Spanish (100%).
The average toxicity score is 0.11.
12 posts have a toxicity score of 0.8 or higher.
Messages with high toxicity are more frequent on Reddit.
About 7% of messages involve Violence.
Keywords frequently found in toxic messages include: coño, joder, puto, negro, and gentuza.

Dutch

This channel contains 2,603 posts, of which 397 are toxic.

Posts were collected between May 08, 2025 and May 20, 2025.
Most data was collected from Twitter (95%), Facebook (2%) and/or Tiktok (2%).
Posts are all written in Dutch (100%).
The average toxicity score is 0.13.
7 posts have a toxicity score of 0.8 or higher.
Messages with high toxicity are more frequent on Facebook.
Toxicity often involves Politics (58%).
About 3% of messages involve Violence.
Keywords frequently found in toxic messages include: freakshow, walgelijk, freak, achterlijke, and uitroeien.

Italian

This channel contains 2,792 posts, of which 156 are toxic.

Posts were collected between May 08, 2025 and May 20, 2025.
Most data was collected from Twitter (96%), Tiktok (3%) and/or Reddit (1%).
Posts are all written in Italian (100%).
The average toxicity score is 0.05.
0 posts have a toxicity score of 0.8 or higher.
Messages with high toxicity are more frequent on TikTok.
Toxicity often involves Profanity (50%).
About 3% of messages involve Violence.
Keywords frequently found in toxic messages include: di merda, coglioni, mannaggia, fanculo, and al rogo.

Voices of division

The online conversation around Eurovision 2025 revealed a layered and regionally varied digital landscape. Overall, toxicity was present in every language we monitored, but its expression and intensity differed. French-speaking content stood out with the highest levels of toxicity, often carrying racially charged rhetoric. In contrast, English and Dutch conversations, though less toxic on average, were saturated with political hostility. These patterns underline Eurovision’s growing role not just as a cultural event but as a flashpoint for ideological expression and identity debates.

Word clouds across all language channels reveal that online Eurovision discourse was heavily shaped by geopolitical conflict. In most languages, terms like Israel and genocide dominated toxic conversations. The German channel stood out, featuring Hamas, antisemitism, and Gaza as the most frequently mentioned terms, suggesting a different framing of the conflict compared to other languages. In Spanish, the most cited term in our dataset was genocidio, indicating a strong emphasis on that specific accusation. By contrast, the Italian word cloud contained hardly any references to the Middle East conflict—apart from ebrei (Jews)—pointing to a less politicised or differently focused conversation.

Platform-specific dynamics also played a role. While mainstream platforms like Twitter (X) dominated in volume, fringe spaces like 4chan, Reddit, and Facebook hosted a disproportionate share of the most extreme posts. The findings point not to a singular narrative, but to a fragmented one where Eurovision becomes a mirror reflecting broader socio-political divides across Europe.

Key Takeaways

1. French-language posts were the most toxic overall

• With an average toxicity score of 0.19 and 823 toxic posts, French was the most toxic language channel in your dataset.

• Racism was the most dominant theme (64%), an outlier compared to other languages, where politics-related discussions were more common.

• This might suggest that Eurovision conversations in French-speaking spaces were particularly racially charged, likely tied to geopolitical tensions (e.g., terms like Israhell).

2. Toxicity was politicised in multiple languages, especially English and Dutch

• Politics was the main theme in toxic English (60%) and Dutch (58%) posts, with terms like whore, freakshow, and walgelijk appearing often.

• In English posts, 60% of the toxic content was political, despite a relatively low average toxicity (0.10).

• This reflects how Eurovision is increasingly politicised, particularly in Anglophone and Dutch-speaking contexts.

3. Platform dynamics differ by language

• 4chan was a key source of high-toxicity posts in English.

• YouTube showed more severe toxicity in German.

• Facebook was most toxic for Dutch-language posts.

• TikTok was the main source of high toxicity in Italian posts, despite the overall low average toxicity there (0.05).

• These variations highlight how language, platform, and community culture intersect to shape online toxicity.

4. Italian conversations stood out for being the least toxic

• With an average toxicity of just 0.05 and no posts scoring 0.8 or higher, Italian discussions were the most civil.

• Toxic messages mainly involved profanity (50%) rather than violent or political rhetoric.

• This may suggest that Italian Eurovision discourse is more emotional or expressive, but less ideologically aggressive.

5. High-toxicity posts were rare but spread across all languages

• Posts with a toxicity score of 0.8 or higher were relatively low in number but appeared in all languages except Italian.

• These messages were more common on fringe or semi-anonymous platforms, such as Reddit and 4chan.

• Their presence across all channels shows that extreme online hate is not limited to one language or region; it’s a pan-European phenomenon.

6. Geopolitical framing varied across languages

• In most language channels, toxic Eurovision conversations were dominated by terms like Israel and genocide, reflecting global tensions spilling into cultural discourse.

• The German-language channel was an exception, where Hamas, antisemitism, and Gaza were the most frequently mentioned terms. The Spanish channel, meanwhile, was strongly centred around the term ‘genocidio’.

• This suggests that different language communities may frame the same controversy in notably different ways, influenced by national media, political culture, and online ecosystems.

From Ballads to Backlash: online toxicity surrounding Eurovision 2025