# Audio Is the Achilles' Heel: Red Teaming Audio Large Multimodal Models

arXiv: https://arxiv.org/abs/2410.23861 · HTML: https://mlsanity.com/paper/2410.23861/

## Key facts

- Cited by 37 papers in the map (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 22 of its 35 references that are in the map.
- First posted on arXiv 31 Oct 2024.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)
- LLM safety and security / [Audio jailbreaks and speech backdoors](https://mlsanity.com/safety/audio-jailbreaks-and-speech-backdoors/)

## Builds on (most cited first, 22 of 22)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal](https://mlsanity.com/paper/2402.04249/) (2024)
- [Jailbreaking Black Box Large Language Models in Twenty Queries](https://mlsanity.com/paper/2310.08419/) (2023)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models](https://mlsanity.com/paper/2311.07919/) (2023)
- [How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs](https://mlsanity.com/paper/2401.06373/) (2024)
- [FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts](https://mlsanity.com/paper/2311.05608/) (2023)
- [Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions](https://mlsanity.com/paper/2309.07875/) (2023)
- [DeepInception: Hypnotize Large Language Model to Be Jailbreaker](https://mlsanity.com/paper/2311.03191/) (2023)
- [Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models](https://mlsanity.com/paper/2307.14539/) (2023)
- [Do-Not-Answer: A Dataset for Evaluating Safeguards in LLMs](https://mlsanity.com/paper/2308.13387/) (2023)
- [Jailbreak Vision Language Models via Bi-Modal Adversarial Prompt](https://mlsanity.com/paper/2406.04031/) (2024)
- [ALERT: A Comprehensive Benchmark for Assessing Large Language Models' Safety through Red Teaming](https://mlsanity.com/paper/2404.08676/) (2024)
- [CValues: Measuring the Values of Chinese Large Language Models from Safety to Responsibility](https://mlsanity.com/paper/2307.09705/) (2023)
- [SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model](https://mlsanity.com/paper/2406.12030/) (2024)
- [Cognitive Overload: Jailbreaking Large Language Models with Overloaded Logical Thinking](https://mlsanity.com/paper/2311.09827/) (2023)
- [SafetyPrompts: a Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safety](https://mlsanity.com/paper/2404.05399/) (2024)
- [Self-Guard: Empower the LLM to Safeguard Itself](https://mlsanity.com/paper/2310.15851/) (2023)
- [Walking a Tightrope -- Evaluating Large Language Models in High-Risk Domains](https://mlsanity.com/paper/2311.14966/) (2023)
- [Investigating Pre-trained Audio Encoders in the Low-Resource Condition](https://mlsanity.com/paper/2305.17733/) (2023)

## Built on by (most cited first, 30 of 37)

- [Best-of-N Jailbreaking](https://mlsanity.com/paper/2412.03556/) (2024)
- [Jailbreak Attacks and Defenses against Multimodal Generative Models: A Survey](https://mlsanity.com/paper/2411.09259/) (2024)
- [Multilingual and Multi-Accent Jailbreaking of Audio LLMs](https://mlsanity.com/paper/2504.01094/) (2025)
- [Jailbreak-AudioBench: In-Depth Evaluation and Analysis of Jailbreak Threats for Large Audio Language Models](https://mlsanity.com/paper/2501.13772/) (2025)
- [Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models](https://mlsanity.com/paper/2505.15406/) (2025)
- [Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models](https://mlsanity.com/paper/2508.07173/) (2025)
- [JALMBench: Benchmarking Jailbreak Vulnerabilities in Audio Language Models](https://mlsanity.com/paper/2505.17568/) (2025)
- [Speech-Audio Compositional Attacks on Multimodal LLMs and Their Mitigation with SALMONN-Guard](https://mlsanity.com/paper/2511.10222/) (2025)
- ["I am bad": Interpreting Stealthy, Universal and Robust Audio Jailbreaks in Audio-Language Models](https://mlsanity.com/paper/2502.00718/) (2025)
- [SARSteer: Safeguarding Large Audio-Language Models via Safe-Ablated Refusal Steering](https://mlsanity.com/paper/2510.17633/) (2025)
- [Reshaping Representation Space to Balance the Safety and Over-rejection in Large Audio Language Models](https://mlsanity.com/paper/2505.19670/) (2025)
- [Universal Acoustic Adversarial Attacks for Flexible Control of Speech-LLMs](https://mlsanity.com/paper/2505.14286/) (2025)
- [SAKE: Towards Editing Auditory Attribute Knowledge of Large Audio-Language Models](https://mlsanity.com/paper/2510.16917/) (2025)
- [Investigating Safety Vulnerabilities of Large Audio-Language Models Under Speaker Emotional Variations](https://mlsanity.com/paper/2510.16893/) (2025)
- [Adversarial Attacks on Multimodal Large Language Models: A Comprehensive Survey](https://mlsanity.com/paper/2603.27918/) (2026)
- [Hijacking Large Audio-Language Models via Context-Agnostic and Imperceptible Auditory Prompt Injection](https://mlsanity.com/paper/2604.14604/) (2026)
- [Acoustic Interference: A New Paradigm Weaponizing Acoustic Latent Semantic for Universal Jailbreak against Large Audio Language Models](https://mlsanity.com/paper/2605.18168/) (2026)
- [Watch, Listen, Understand, Mislead: Tri-modal Adversarial Attacks on Short Videos for Content Appropriateness Evaluation](https://mlsanity.com/paper/2507.11968/) (2025)
- [ERIS: Evolutionary Real-world Interference Scheme for Jailbreaking Audio Large Models](https://mlsanity.com/paper/2509.11128/) (2025)
- [MUGEN: Evaluating and Improving Multi-audio Understanding of Large Audio-Language Models](https://mlsanity.com/paper/2603.09714/) (2026)
- [Resurfacing Paralinguistic Awareness in Large Audio Language Models](https://mlsanity.com/paper/2603.11947/) (2026)
- [CASA: Classification Augmented with Safety Attention for Robust Multimodal Alignment](https://mlsanity.com/paper/2604.00310/) (2026)
- [GRM: Utility-Aware Jailbreak Attacks on Audio LLMs via Gradient-Ratio Masking](https://mlsanity.com/paper/2604.09222/) (2026)
- [Sparse Tokens Suffice: Jailbreaking Audio Language Models via Token-Aware Gradient Optimization](https://mlsanity.com/paper/2605.04700/) (2026)
- [Investigating Vulnerabilities and Defenses Against Audio-Visual Attacks: A Comprehensive Survey Emphasizing Multimodal Models](https://mlsanity.com/paper/2506.11521/) (2025)
- [Beyond Text: Multimodal Jailbreaking of Vision-Language and Audio Models through Perceptually Simple Transformations](https://mlsanity.com/paper/2510.20223/) (2025)
- [Rebellion: Noise-Robust Reasoning Training for Audio Reasoning Models](https://mlsanity.com/paper/2511.09682/) (2025)
- [The Alignment Curse: Modality Alignment Supercharges Audio Attacks via Text Transfer](https://mlsanity.com/paper/2602.02557/) (2026)
- [Aegis: Towards Governance, Integrity, and Security of AI Voice Agents](https://mlsanity.com/paper/2602.07379/) (2026)
- [Evaluation of Audio Language Models for Fairness, Safety, and Security](https://mlsanity.com/paper/2603.13262/) (2026)

## Abstract (arXiv)

Large Multimodal Models (LMMs) have demonstrated the ability to interact with humans under real-world conditions by combining Large Language Models (LLMs) and modality encoders to align multimodal information (visual and auditory) with text. However, such models raise new safety challenges of whether models that are safety-aligned on text also exhibit consistent safeguards for multimodal inputs. Despite recent safety-alignment research on vision LMMs, the safety of audio LMMs remains under-explored. In this work, we comprehensively red team the safety of five advanced audio LMMs under three settings: (i) harmful questions in both audio and text formats, (ii) harmful questions in text format accompanied by distracting non-speech audio, and (iii) speech-specific jailbreaks. Our results under these settings demonstrate that open-source audio LMMs suffer an average attack success rate of 69.14% on harmful audio questions, and exhibit safety vulnerabilities when distracted with non-speech audio noise. Our speech-specific jailbreaks on Gemini-1.5-Pro achieve an attack success rate of 70.67% on the harmful query benchmark. We provide insights on what could cause these reported safety-misalignments. Warning: this paper contains offensive examples.