Elamur Research⌕
☰
home / papers / 2026 / Toward Dialect-Aware Safety Evaluation for Arabic Large Language Models

Toward Dialect-Aware Safety Evaluation for Arabic Large Language Models

Paper · 2026 · Wajdi Zaghouani

Summary

LLM safety testing focuses mostly on English and standard language forms, yet most Arabic speakers write in regional dialects that alignment data and benchmarks rarely cover. The paper names this problem the Dialect Safety Gap: a model responds differently to the same harmful request depending on the dialect it is phrased in. Drawing on algorithmic fairness, the author formalizes dialect safety consistency and links the gap to unequal false-negative rates across groups. Two metrics measure it. A binary DSG Score gives a coarse view, and a Pairwise Dialect Inconsistency (PDI) matrix captures magnitude and direction. The gap is attributed to lexical, morphological and pragmatic divergence from the norms encoded during alignment. The paper also proposes the Dialect-Aware Safety Evaluation Protocol (DASEP), which combines naturalistically sourced dialect prompts with MSA-pivot baselines. A small human-annotated probe of one model across five Arabic variety groups found structured safety degradation. The best- and worst-protected varieties corresponded to their closeness to and distance from MSA. This matters because the communities most exposed to online harm may get the weakest protection.

Key points

  • The five variety groups are MSA, Egyptian, Levantine, Gulf and Maghrebi, following the grouping in MADAR (Bouamor et al., 2018). Maghrebi is the most distant from MSA, with Berber and French borrowings.
  • The probe used 20 MSA-authored base prompts rendered into 5 varieties, giving 100 instances, all tested on a single model.
  • Responses were scored as safe or unsafe. Both qualified compliance with warnings (W) and full harmful generation (C) counted as unsafe.
  • The author stresses that the results are a proof of concept, not population-level estimates. The study lacks benign prompts (so no false-positive rates), naturalistic sourcing (e.g., from L-HSAB) and normalization experiments.
  • The prompt set is withheld to prevent misuse. The author plans to release code for computing the DSG Score and PDI matrix. The work was funded by QRDI grant NPRP14C-0916-210015.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Northwestern University in QatarDoha · Qatar39 works · 148 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksSocial Sciences · 1,954 worksSafety Research · 626 works

🧭 Topics 4Research subjects and languages the work is about.Language models & pretraining · 3,894 worksSemitic languages · 3,066 worksEvaluation & benchmarks · 1,341 worksBias, fairness & safety · 587 works

Cite

Cite as Wajdi Zaghouani. 2026. Toward Dialect-Aware Safety Evaluation for Arabic Large Language Models. Proceedings of the 6th Workshop on Trustworthy NLP (TrustNLP 2026).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.trustnlp-main.37

@inproceedings{zaghouani2026toward,
  title = {{Toward Dialect-Aware Safety Evaluation for Arabic Large Language Models}},
  author = {Wajdi Zaghouani},
  booktitle = {Proceedings of the 6th Workshop on Trustworthy NLP (TrustNLP 2026)},
  year = {2026},
  doi = {10.18653/v1/2026.trustnlp-main.37},
  url = {https://doi.org/10.18653/v1/2026.trustnlp-main.37}
}

References

37 references, 26 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Muhammad Abdul-Mageed, Chiyu Zhang, Houda Bouamor, and Nizar Habash. 2020. NADI 2020: The first nuanced Arabic dialect identification shared task. In Proceedings of the Fifth Arabic Natural Language Processing Workshop, pages 97–110. Association for Computational Linguistics. ↗
  2. Ibrahim Abu Farha and Walid Magdy. 2020. Multitask learning for Arabic offensive language and hate-speech detection. In Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools, with a Shared Task on Offensive Language Detection, pages 86–90. European Language Resources Association.
  3. Abubakar Abid, Maheen Farooqi, and James Zou. 2021. Persistent anti-Muslim bias in large language models. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, pages 298–306.
  4. Yuntao Bai, Andy Jones, Kamal Ndousse, Amanda Askell, Anna Chen, Nova DasSarma, Dawn Drain, Stanislav Fort, Deep Ganguli, Tom Henighan, Nicholas Joseph, Saurav Kadavath, Jackson Kernion, Tom Conerly, Sheer El-Showk, Nelson Elhage, Zac Hatfield-Dodds, Danny Hernandez, Tristan Hume, Scott Johnston, Shauna Kravec, Liane Lovitt, Neel Nanda, Catherine Olsson, Dario Amodei, Tom Brown, Jack Clark, Sam McCandlish, Chris Olah, Ben Mann, and Jared Kaplan. 2022. Training a helpful and harmless assistant with reinforcement learning from human feedback. arXiv preprint arXiv:2204.05862. ↗
  5. Solon Barocas and Andrew D. Selbst. 2016. Big data’s disparate impact. California Law Review, 104(3):671–732. find
  6. Su Lin Blodgett, Solon Barocas, Hal Daumé III, and Hanna Wallach. 2020. Language (technology) is power: A critical survey of “bias” in NLP. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5454– 5476.
  7. Rishi Bommasani, Drew A. Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S. Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, Erik Brynjolfsson, Shyamal Buch, Dallas Card, Rodrigo Castellon, Niladri Chatterji, Annie Chen, Kathleen Creel, Jared Quincy Davis, Dora Demszky, Chris Donahue, Moussa Doumbouya, Esin Durmus, Stefano Ermon, John Etchemendy, Kawin Ethayarajh, Li Fei-Fei, Chelsea Finn, Trevor Gale, Lauren Gillespie, Karan Goel, Noah Goodman, Shelby Grossman, Neel Guha, Tatsunori Hashimoto, Peter Henderson, John Hewitt, Daniel E. Ho, Jenny Hong, Kyle Hsu, Jing Huang, Thomas Icard, Saahil Jain, Dan Jurafsky, Pratyusha Kalluri, Siddharth Karamcheti, Geoff Keeling, Fereshte Khani, Omar Khattab, Pang Wei Koh, Mark Krass, Ranjay Krishna, Rohith Kuditipudi, and Percy Liang. 2021. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258. ↗
  8. Houda Bouamor, Nizar Habash, Mohammad Salameh, Wajdi Zaghouani, Owen Rambow, Dana Abdulrahim, Ossama Obeid, Salam Khalifa, Fadhl Eryani, Alexander Erdmann, and Kemal Oflazer. 2018. The MADAR Arabic dialect corpus and lexicon. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), pages 3387–3396. European Language Resources Association. find
  9. Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. Unsupervised cross-lingual representation learning at scale. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8440– 8451.
  10. Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia Gonzalez, Prangthip Hansanti, John Hoffman, Semarley Jarrett, Kaushik Ram Sadagopan, Dirk Rowe, Shannon Spruit, Chau Tran, Pierre Andrews, Necip Fazil Ayan, Shruti Bhosale, Sergey Edunov, Angela Fan, Cynthia Gao, Vedanuj Goswami, Francisco Guzmán, Philipp Koehn, Alexandre Mourachko, Christophe Ropers, Safiyyah Saleem, Holger Schwenk, and Jeff Wang. 2022. No language left behind: Scaling human-centered machine translation. arXiv preprint arXiv:2207.04672. ↗
  11. Kareem Darwish, Nizar Habash, Mourad Abbas, Hend Al-Khalifa, Huseein T. Al-Natsheh, Samhaa R. El-Beltagy, Houda Bouamor, Karim Bouzoubaa, Violetta Cavalli-Sforza, Wassim El-Hajj, Mustafa Jarrar, and Hamdy Mubarak. 2021. A panoramic survey of natural language processing in the Arab world. Communications of the ACM, 64(4):72–81. ↗
  12. Thomas Davidson, Dana Warmsley, Michael Macy, and Ingmar Weber. 2017. Automated hate speech detection and the problem of offensive language. In Proceedings of the Eleventh International AAAI Conference on Web and Social Media, pages 512–515. ↗
  13. Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, and Richard Zemel. 2012. Fairness through awareness. In Proceedings of the 3rd Innovations in Theoretical Computer Science Conference, pages 214–226. ↗
  14. Charles A. Ferguson. 1959. Diglossia. Word, 15(2):325– 340. find
  15. Deep Ganguli, Liane Lovitt, Jackson Kernion, Amanda Askell, Yuntao Bai, Saurav Kadavath, Ben Mann, Ethan Perez, Nicholas Schiefer, Kamal Ndousse, Andy Jones, Sam Bowman, Anna Chen, Tom Conerly, Nova DasSarma, Dawn Drain, Nelson Elhage, Sheer El-Showk, Stanislav Fort, Zachary Dodds, Tom Henighan, Danny Hernandez, Tristan Hume, Josh Jacobson, Scott Johnston, Shauna Kravec, Catherine Olsson, Sam Ringer, Eli Tran-Johnson, Dario Amodei, Tom Brown, Nicholas Joseph, Sam McCandlish, Chris Olah, Jared Kaplan, and Jack Clark. 2022. Red teaming language models to reduce harms: Methods, scaling behaviors, and lessons learned. arXiv preprint arXiv:2209.07858. ↗
  16. Samuel Gehman, Suchin Gururangan, Maarten Sap, Yejin Choi, and Noah A. Smith. 2020. RealToxicityPrompts: Evaluating neural toxic degeneration in language models. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3356–3369.
  17. Nizar Y. Habash. 2010. Introduction to Arabic Natural Language Processing. Morgan & Claypool Publishers. ↗
  18. Thomas Hartvigsen, Saadia Gabriel, Hamid Palangi, Maarten Sap, Dipankar Ray, and Ece Kamar. 2022. ToxiGen: A large-scale machine-generated dataset for adversarial and implicit hate speech detection. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3309–3326.
  19. Dan Hendrycks, Collin Burns, Steven Basart, Andrew Critch, Jerry Li, Dawn Song, and Jacob Steinhardt. 2021. Aligning AI with shared human values. arXiv preprint arXiv:2008.02275. ↗
  20. Sharon Levy, Emily Allaway, Melanie Subbiah, Lydia Chilton, Desmond Patton, Kathleen McKeown, and William Yang Wang. 2022. SafeText: A benchmark for exploring physical safety in language models. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2407–2421. ↗
  21. Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, Benjamin Newman, Binhang Yuan, Bobby Yan, Ce Zhang, Christian Cosgrove, Christopher D. Manning, Christopher Ré, Diana Acosta-Navas, Drew A. Hudson, Eric Zelikman, Esin Durmus, Faisal Ladhak, Frieda Rong, Hongyu Ren, Huaxiu Yao, Jue Wang, Keshav Santhanam, Laurel Orr, Lucia Zheng, Mert Yuksekgonul, Mirac Suzgun, Nathan Kim, Neel Guha, Niladri Chatterji, Omar Khattab, Peter Henderson, Qian Huang, Ryan Chi, Sang Michael Xie, Shibani Santurkar, Surya Ganguli, Tatsunori Hashimoto, Thomas Icard, Tianyi Zhang, Vishrav Chaudhary, William Wang, Xuechen Li, Yifan Mai, Yuhui Zhang, and Yuta Bisk. 2022. Holistic evaluation of language models. arXiv preprint arXiv:2211.09110. ↗
  22. Stephanie Lin, Jacob Hilton, and Owain Evans. 2022. TruthfulQA: Measuring how models mimic human falsehoods. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3214–3252.
  23. Hala Mulki, Hatem Haddad, Chedi Bechikh Ali, and Halima Alshabani. 2019. L-HSAB: A Levantine Twitter dataset for hate speech and abusive language. In Proceedings of the Third Workshop on Abusive Language Online, pages 111–118.
  24. Hamdy Mubarak, Kareem Darwish, and Walid Magdy. 2017. Abusive language detection on Arabic social media. In Proceedings of the First Workshop on Abusive Language Online, pages 52–56.
  25. Nayla Ousidhoum, Zizheng Lin, Hongming Zhang, Yangqiu Song, and Dit-Yan Yeung. 2019. Multilingual and multi-aspect hate speech analysis. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 4675–4684.
  26. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul Christiano, Jan Leike, and Ryan Lowe. 2022. Training language models to follow instructions with human feedback. In Advances in Neural Information Processing Systems, volume 35, pages 27730–27744.
  27. Laura Weidinger, Jonathan Uesato, Maribeth Rauh, Conor Griffin, Po-Sen Huang, John Mellor, Amelia Glaese, Myra Cheng, Borja Balle, Atoosa Kasirzadeh, Courtney Biles, Sasha Brown, Zac Kenton, Will Hawkins, Tom Stepleton, Abeba Birhane, Lisa Anne Hendricks, Laura Rimell, William Isaac, Julia Haas, Sean Legassick, Geoffrey Irving, and Iason Gabriel. 2022. Taxonomy of risks posed by language models. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, pages 214–229. find
  28. Zheng-Xin Yong, Cristina Menghini, and Stephen H. Bach. 2023. Low-resource languages jailbreak GPT-4. arXiv preprint arXiv:2310.02446. ↗
  29. Ethan Perez, Saffron Huang, Francis Song, Trevor Cai, Roman Ring, John Aslanides, Amelia Glaese, Nat McAleese, and Geoffrey Irving. 2022. Red teaming language models with language models. arXiv preprint arXiv:2202.03286. ↗
  30. Wajdi Zaghouani, Behrang Mohit, Nizar Habash, Ossama Obeid, Nadi Tomeh, Alla Rozovskaya, Noura Farra, Sarah Alkuhlani, and Kemal Oflazer. 2014. Large scale Arabic error annotation: Guidelines and framework. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC 2014), pages 2362–2369. ↗
  31. Paul Röttger, Bertie Vidgen, Dong Nguyen, Zeerak Talat, Helen Margetts, and Janet Pierrehumbert. 2021. HateCheck: Functional tests for hate speech detection models. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 41–58. ↗
  32. Marcos Zampieri, Shervin Malmasi, Preslav Nakov, Sara Rosenthal, Noura Farra, and Ritesh Kumar. 2019. Predicting the type and target of offensive posts in social media. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 1415–1420.
  33. Mohammad Salameh, Houda Bouamor, and Nizar Habash. 2018. Fine-grained Arabic dialect identification. In Proceedings of the 27th International Conference on Computational Linguistics, pages 1332– 1344.
  34. Tony Sun, Andrew Gaut, Shirlyn Tang, Yuxin Huang, Mai ElSherief, Jieyu Zhao, Diba Mirza, Elizabeth Belding, Kai-Wei Chang, and William Yang Wang. 2019. Mitigating gender bias in natural language processing: Literature review. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1630–1641. ↗
  35. Yuxia Wang, Haonan Li, Xudong Han, Preslav Nakov, and Timothy Baldwin. 2023. Do-not-answer: A dataset for evaluating safeguards in LLMs. arXiv preprint arXiv:2308.13387. ↗
  36. Zeerak Waseem and Dirk Hovy. 2016. Hateful symbols or hateful people? Predictive features for hate speech detection on Twitter. In Proceedings of the NAACL Student Research Workshop, pages 88–93.
  37. Laura Weidinger, John Mellor, Maribeth Rauh, Conor Griffin, Jonathan Uesato, Po-Sen Huang, Myra Cheng, Mia Glaese, Borja Balle, Atoosa Kasirzadeh, Zac Kenton, Sasha Brown, Will Hawkins, Tom Stepleton, Courtney Biles, Abeba Birhane, Julia Haas, Laura Rimell, Lisa Anne Hendricks, William Isaac, Sean Legassick, Geoffrey Irving, and Iason Gabriel. 2021. Ethical and social risks of harm from language models. arXiv preprint arXiv:2112.04359. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zaghouani-2026-dialect-aware-safety-evaluation-arabic/