Lune

ACL Findings2026

SEA-Guard: Culturally Grounded Multilingual Safeguard for Southeast Asia

Panuthep Tasawong, Jian Gang Ngui, Alham Fikri Aji, Trevor Cohn, Peerat Limkonchotiwat

Abstract

Culturally aware safeguards are crucial for AI alignment in real-world settings, where safety extends beyond common sense and encompasses diverse local values, norms, and regionspecific regulations. However, building largescale, culturally grounded datasets is challenging due to limited resources and a scarcity of native annotators. Consequently, many safeguard models rely on machine translation of English datasets, often missing regional and cultural nuances. We present a novel agentic data-generation framework to scalably create authentic, region-specific safety datasets for Southeast Asia (SEA). On this foundation, we introduce the SEA-Guard family, the first multilingual safeguard models grounded in SEA cultural contexts. Evaluated across multiple benchmarks and cultural variants, SEA-Guard consistently outperforms existing safeguards at detecting regionally sensitive or harmful content while maintaining strong general safety performance.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

1 top-tier paper has cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Building Benchmarks from the Ground Up: Community-Centered Evaluation of LLMs in Healthcare Chatbot Settings cites CultureLLM: Incorporating Cultural Differences into Large Language Models
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers1

  1. Building Benchmarks from the Ground Up: Community-Centered Evaluation of LLMs in Healthcare Chatbot SettingsHamna, Gayatri Bhat, Sourabrata Mukherjee, Faisal M. Lalani et al.CHI 2026 · 1 citation

Builds on16

  1. Chain-of-Thought Prompting Elicits Reasoning in Large Language ModelsJason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma et al.NeurIPS 2022 · 22,562 citations
  2. WildTeaming at Scale: From In-the-Wild Jailbreaks to (Adversarially) Safer Language ModelsLiwei Jiang, Kavel Rao, Seungju Han, Allyson Ettinger et al.NeurIPS 2024 · 247 citations
  3. Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language ModelsYongshuo Zong, Ondrej Bohdal, Tingyang Yu, Yongxin Yang et al.ICML 2024 · 140 citations
  4. CultureLLM: Incorporating Cultural Differences into Large Language ModelsCheng Li, Mengzhuo Chen, Jindong Wang, Sunayana Sitaram et al.NeurIPS 2024 · 101 citations
  5. Best-of-N JailbreakingJohn Hughes, Sara Price, Aengus Lynch, Rylan Schaeffer et al.NeurIPS 2025 · 78 citations

Related papers

  1. SEA-SafeguardBench: Culturally Grounded Safety Benchmark for Southeast Asian LanguagesPanuthep Tasawong, Jian Gang Ngui, Alham Fikri Aji, Trevor Cohn et al.ACL Findings 2026
  2. Toxicity Red-Teaming: Benchmarking LLM Safety in Singapore's Low-Resource LanguagesYujia Hu, Ming Shan Hee, Preslav Nakov, Roy Ka-Wei LeeEMNLP 2025
  3. MrGuard: A Multilingual Reasoning Guardrail for Universal LLM SafetyYahan Yang, Soham Dan, Shuo Li, Dan Roth et al.EMNLP 2025
  4. UbuntuGuard: A Culturally-Grounded Policy Benchmark for Equitable AI Safety in African LanguagesTassallah Abdullahi, Macton Mgonzo, Mardiyyah Oduwole, Paul Okewunmi et al.ACL Findings 2026
  5. Crowdsource, Crawl, or Generate? Creating SEA-VL, a Multicultural Vision-Language Dataset for Southeast AsiaSamuel Cahyawijaya, Holy Lovenia, Joel Ruben Antony Moniz, Tack Hwa Wong et al.ACL 2025
Dusk over the sea between two cliffs drawn in fine vertical lines