TextDetox CLEF 2025: Join the Multilingual Text Detoxification Challenge
Join the TextDetox CLEF 2025 Shared Task to develop innovative methods for transforming harmful language into neutral or constructive text while preserving meaning.
Calling all NLP researchers and practitioners! The TextDetox CLEF 2025 Shared Task invites you to develop innovative methods for transforming harmful language into neutral or constructive text while preserving meaning.
This multilingual challenge spans 15 languages, pushing the boundaries of responsible AI and ethical NLP. Whether you’re an expert in text style transfer, toxicity detection, or multilingual NLP, this is your chance to contribute to a safer, more inclusive digital space!
Learn more about the challenge and explore the TextDetox dataset.
Key highlights:
- Transfer a text style from toxic to neutral (e.g., ‘what a f**k is this about?’ -> ‘what is this about?’)
- 15 diverse languages: English, Spanish, Italian, French, Chinese, Japanese, Hindi, Hinglish, Arabic, German, Russian, Ukrainian, Amharic, Hebrew, Tatar
- Two phases: dev and test with more data to challenge your models!
Important dates:
- Now: join our Google group for updates (https://groups.google.com/g/textdetox-clef2025)
- April 25, 2025: Registration closes
- May 1st, 2025: Test phase starts
- May 10, 2025: End of evaluation cycle
- May 30, 2025: Participants’ paper submission
- June 27, 2025: Notification of acceptance
- July 7, 2025: Camera-ready due
- September 9-12, 2025: CLEF Conference in Madrid, Spain!
Tags: TextDetox CLEF 2025, Multilingual Text Detoxification, NLP, Responsible AI, Ethical NLP, Text Style Transfer, Toxicity Detection, Multilingual NLP