{
  "id": 4757559,
  "title": "Not Safe for All: Auditing the Dialect Penalty in Text-to-Image Safety Pipelines",
  "url": "https://urgent.news/2026/08/30/not-safe-for-all-auditing-the-dialect-penalty-in-text-to-image-safety",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-30T06:19:46.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2608.29589v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "Text-to-image (T2I) safety guardrails fail to generalize equitably to non-standard dialects. Evaluating 23,080 paired prompts across five English dialects, we formalize this failure as the dialect penalty, where filters trigger based on linguistic surface features rather than semantic intent. Text-level filters fail in opposing directions: NSFW-T over-flags benign dialect prompts and LatentGuard…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}