{
  "id": 18429,
  "url": "https://arxiv.org/abs/2608.10405v1",
  "title": "Never Stop Speaking: a Denial-of-Service Attack on End-to-End Speech Language Models",
  "summary": "Many studies have shown that specially crafted inputs can induce large language models (LLMs) to generate excessively long outputs, resulting in significant computational overhead and resource consumption. While most existing denial-of-service (DoS) attacks target text-only LLMs, end-to-end (E2E) speech LLMs are rapidly emerging. Existing text-based DoS attacks primarily rely on prompt engineering, such as adversarial suffixes or semantic inducement, which exploit the discrete nature of text inp",
  "authors": "Shuozhe Cheng, Kunlan Xiang, Mingxuan Li, Ji Zhang, Dongxiao Liu, Wenbo Jiang",
  "category": "research",
  "topics": null,
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-11T02:50:30.000Z",
  "fetched_at": "2026-08-12T05:10:43.828Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/18429",
  "original_url": "https://arxiv.org/abs/2608.10405v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}