Greetings All,
We are delighted to announce SemEval-2026 Task 9: Detecting Multilingual, Multicultural, and Multievent Online Polarization (POLAR@2026).
Important Updates
The development phase deadline has been extended to 24 January, 23:59 (UTC-12)
The evaluation phase will run from 25 January to 2 February (UTC-12)
New registrations will close on 24 January
Virtual Seminar
We will host a virtual seminar on 23 January (10:00–11:00 CET), covering:
Task overview and datasets
Technical details on Codabench
Open Q&A session with the organisers
———————
Languages
———————
We provide data in 20+ languages, including: German, Spanish, English, Arabic, Hausa, Urdu, Amharic, Italian, Russian, Myanmar, Chinese, Nepali, Hindi, Telugu, Persian, Turkish, Bangla, Somali, Emakhuwa, Mozambican Portuguese, Igbo, Khmer, Odia, and Punjabi.
———————
Subtasks
———————
Subtask 1 (LIVE): Polarization Detection – Identify whether a text exhibits polarization.
https://www.codabench.org/competitions/10522/
Subtask 2: Polarization Type Classification – Classify polarized content into specific types (e.g., political, social, cultural).
https://www.codabench.org/competitions/10669/
Subtask 3: Manifestation Identification – Determine how polarization is expressed (e.g., language cues, tone, argumentative structure).
https://www.codabench.org/competitions/10674/
———————
Participation & Resources
———————
Task Website:
https://polar-semeval.github.io/
Discord (community Q&A and discussion):
https://discord.gg/K5uhEF2jPN
Google Group (official updates & registration):
https://groups.google.com/g/polarization-semeval-2026-participants
We warmly invite the community to participate in this shared task and contribute to advancing multilingual, multicultural, and socially grounded NLP research on online polarization.
Best regards,
The SemEval-2026 Task 9 Organizers