Call for Participation - SemEval-2026 Task 9: Detecting Multilingual, Multicultural, and Multievent Online Polarization

[Shamsuddeen Muhammad]

Dear colleagues,

We are pleased to announce SemEval-2026 Task 9: Detecting Multilingual, Multicultural, and Multievent Online Polarization. This shared task aims to advance understanding of how polarization appears in text across different languages, cultures, and events. Participants will develop models to detect and interpret polarized content in contexts such as elections, conflicts, protests, and public debates.

Languages (20+): 

German, Spanish, English, Arabic, Hausa, Urdu, Amharic, Italian, Russian, Myanmar, Chinese, Nepali, Hindi, Telugu, Persian, Turkish, Bangla, Somali, Emakhuwa, Mozambican Portuguese, Igbo (Nigeria), Khmer, Odia, and Punjabi.

Subtasks

• Subtask 1: Polarization Detection – Identify whether a text exhibits polarization.
• Subtask 2: Polarization Type Classification – Classify polarized content into specific types (e.g., political, social, cultural).
• Subtask 3: Manifestation Identification – Determine how polarization is expressed or manifested (e.g., linguistic cues, tone, argumentative structure).

Task Page: https://polar-semeval.github.io/.  

Discord (community Q&A and discussion): https://discord.gg/RF9MzYwKZP

On behalf of the SemEval-2026 Task 9 Organizer