# Preprint introduces Arabic-language LLM safety benchmark

_Published Tuesday, August 25, 2026 at 9:07 PM EDT · Science · Latest · Tier 2 — Notable_

A preprint introduces the Arabic Safety Index, a human-curated red-teaming benchmark with 801 prompts across eight safety categories and eight attack strategies. Its authors evaluated seven Arabic-capable models and report that most failed to defend against 50% of unsafe prompts. The study also reports that direct and obfuscation-based attacks were most effective and that automated safety judges performed poorly against human annotators.

## Sources

- [cs.AI updates on arXiv.org](https://arxiv.org/abs/2608.21985)
- [cs.AI updates on arXiv.org](https://arxiv.org/abs/2604.01532)

---
Canonical: https://techandbusiness.org/newswire/V9A12W2Ud4PAabRcEVgrrk
Published: 2026-08-26T01:07:46.417Z
Story chronology: 2026-08-25T04:00:00.000Z
Retrieved: 2026-10-10T11:00:32.947Z
Publisher: Tech & Business (techandbusiness.org)