Multimodal Hateful Meme Detection
Detect hateful Arabic memes from image + extracted text.
- Dev dataset released with labels for local testing
- Submit predictions on CodaBench (subtasks A1 & A2)
ArabicNLP 2026 Shared Task
ArGuard
Harmful Content Detection in Arabic Memes and LLM Prompts
A two-track benchmark for Arabic content safety research, combining multimodal hateful meme understanding with harmful prompt detection for LLM interactions.
Overview
Arabic safety evaluation across modalities
Harmful Arabic content appears in multiple forms: memes that combine visual cues with embedded text, and prompts that seek unsafe responses from large language models. ArGuard evaluates both settings with a shared design that moves from binary detection to fine-grained categorization.
The task emphasizes Arabic-specific challenges such as dialect use, sarcasm, code-switching, cultural references, and the interaction between language and visual context.
Live updates
What’s happening right now
Latest releases and announcements from the organizers.
Harmful Prompt Detection
Detect unsafe Arabic prompts directed at LLMs.
- Dev dataset released with labels for local testing
- External datasets allowed for training (e.g., AraSafe — Benchmarking Safety in Arabic LLMs)
ArGuard Task 1 dataset is live
Training, development, and dev-test splits are available on the Hugging Face Hub.
- QCRI/ArGuard-Task1 is publicly available
- Dev split ships with gold labels for local evaluation
Tasks
Two complementary tracks
Participants may work on either track or both. Each track includes a binary classification subtask and a fine-grained categorization subtask.
Track A
Multimodal Hateful Meme Detection
Systems receive an Arabic meme image with extracted text and predict whether the meme is hateful, then identify the fine-grained label set associated with the content.
- A1: Hateful vs. Not Hateful classification.
- A2: Multi-label fine-grained category prediction.
Labels include mocking, incitement, dehumanization, slurs, contempt, inferiority, exclusion, stereotyping, extremism, threat, insults, historical, humor, sarcasm, and other categories.
Task DetailsTrack B
Textual Harmful Prompt Detection
Systems receive an Arabic prompt directed at an LLM and decide whether it is safe or unsafe, then classify unsafe prompts by harm domain.
- B1: Safe vs. Unsafe prompt detection.
- B2: Harm category classification.
Harm domains include self-harm, harm to others, harassment, adult content, bullying, hate speech, and fraud or illegal activities.
Task DetailsTimeline
Important dates
All dates are tentative and aligned with the ArabicNLP 2026 shared task schedule.
- Registration deadline and beginning of the evaluation cycle (test sets release).
- End of the evaluation cycle (run submission).
- Release leaderboard.
- Shared task papers due date.
- Notification of acceptance.
- Camera-ready papers due.
- ArabicNLP conference.
Resources
Core participant resources will be updated here as data, starter-kit material, baselines, and the CodaBench competition become available.
Organizers
Task organizing team
Firoj Alam
Qatar Computing Research Institute, HBKU
Md. Rafiul Biswas
Hamad Bin Khalifa University
Mohamed Bayan Kmainasi
Qatar Computing Research Institute, HBKU
Ali Ezzat Shahroor
Qatar Computing Research Institute, HBKU
Hamdy Mubarak
Qatar Computing Research Institute, HBKU
Georgios Mikros
Hamad Bin Khalifa University, Qatar
Abul Hasnat
APAVI.AI, France
Wajdi Zaghouani
Northwestern University in Qatar
Acknowledgments
Acknowledgments
The contributions of this work were funded by the NPRP grant 14C-0916-210015, which is provided by the Qatar National Research Fund (a member of Qatar Foundation).
Contact
Questions?
For task updates, questions, and coordination, contact the organizers by email or slack.