mlsanityData to 14 Oct 2026

ARENA: Automated Red-Teaming for Large Audio Language Models

arXiv 2608.15578 ↗First posted 16 Aug 2026cs.SD cs.AIcited by 0

On the map: LLM safety and security → Audio jailbreaks and speech backdoors

Abstract (arXiv)

Large audio-language models (LALMs) make it possible to interact with language models through speech, music, and environmental sound, but they also introduce a safety surface that is difficult to expose with text-only red-teaming. We study automated audio-grounded red-teaming, where a text query must remain safe in isolation while the joint text-audio input induces harmful target behavior. We propose ARENA, a closed-loop framework that trains a controller on an independent 2,000case text-audio dataset. MD-Judge supplies training rewards and adaptive search feedback, while a separate, non-adaptive Llama Guard 3 evaluator alone labels final outcomes. On 520 held-out AdvBench objectives, ARENA achieves FDR/PSR of 87.9/100.0%, 71.5/96.3%, 68.1/100.0%, and 75.4/98.5% on Audio Flamingo 3, Qwen2-Audio, MiMo-Audio, and GPTAudio, respectively. Ablations show that feedback-based refinement and audio-variant search substantially improve attack discovery.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Audio jailbreaks and speech backdoors, one of 12 directions in LLM safety and security. Builds on 21 of its 30 references that are in the map. First posted on arXiv 16 Aug 2026.

Builds on

21 of 30 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025Speech-Audio Compositional Attacks on Multimodal LLMs and Their Mitigation with SALMONN-Guard13 Nov

Other references · 20

  1. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  2. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
  3. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  4. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,429
  5. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
  6. 2022Red Teaming Language Models with Language Modelscited by 1,026
  7. 2022Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentationcited by 719
  8. 2023Tree of Attacks: Jailbreaking Black-Box LLMs Automaticallycited by 672
  9. 2023GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Promptscited by 560
  10. 2023SpeechGPT: Empowering Large Language Models with Intrinsic Cross-Modal Conversational Abilitiescited by 529
  11. 2023AudioPaLM: A Large Language Model That Can Speak and Listencited by 343
  12. 2023Pengi: An Audio Language Model for Audio Taskscited by 220
8 more
  1. 2023MMA-Diffusion: MultiModal Attack on Diffusion Modelscited by 194
  2. 2023SneakyPrompt: Jailbreaking Text-to-image Generative Modelscited by 189
  3. 2024Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilitiescited by 186
  4. 2024ART: Automatic Red-teaming for Text-to-Image Models to Protect Benign Userscited by 35
  5. 2024Groot: Adversarial Testing for Generative Text-to-Image Models with Tree-based Semantic Transformationcited by 14
  6. 2025Red-Teaming Text-to-Image Systems by Rule-based Preference Modelingcited by 8
  7. 2025Investigating Safety Vulnerabilities of Large Audio-Language Models Under Speaker Emotional Variationscited by 4
  8. 2025TEAR: Temporal-aware Automated Red-teaming for Text-to-Video Modelscited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Audio Is the Achilles' Heel: Red Teaming Audio Large Multimodal Models
  2. 2026A Survey of Large Audio Language Models: Generalization, Trustworthiness, and Outlook
  3. 2026Hijacking Large Audio-Language Models via Context-Agnostic and Imperceptible Auditory Prompt Injection
  4. 2026Audio Jailbreaks in Large Audio-Language Models: Taxonomy, Attack-Defense Analysis, and Cost-Aware Evaluation
  5. 2026Red-TTT: Test-Time Training for Automated Jailbreaking Large Language Models
  6. 2025When Good Sounds Go Adversarial: Jailbreaking Audio-Language Models with Benign Inputs
  7. 2025`From Prompt to Perturbation': An Adaptive Framework for Voice-Based Jailbreaks on Audio LLMs
  8. 2025Learning-Based Automated Adversarial Red-Teaming for Robustness Evaluation of Large Language Models