StyleBreak: Revealing Alignment Vulnerabilities in Large Audio-Language Models via Style-Aware Audio Jailbreak
Preprint
Abstract
Large Audio-language Models (LAMs) have recently enabled powerful speech-based interactions by coupling audio encoders with Large Language Models (LLMs). However, the security of LAMs under adversarial attacks remains underexplored, especially through audio jailbreaks that craft malicious audio prompts to bypass alignment. Existing efforts primarily rely on converting text-based attacks into speech or applying shallow signal-level perturbations, overlooking the impact of human speech's expressive variations on LAM alignment robustness. To address this gap, we propose StyleBreak, a novel style-aware audio jailbreak framework that systematically investigates how diverse human speech attributes affect LAM alignment robustness. Specifically, StyleBreak employs a two-stage style-aware transformation pipeline that perturbs both textual content and audio to control linguistic, paralinguistic, and extralinguistic attributes. Furthermore, we develop a query-adaptive policy network that automatically searches for adversarial styles to enhance the efficiency of LAM jailbreak exploration. Extensive evaluations demonstrate that LAMs exhibit critical vulnerabilities when exposed to diverse human speech attributes. Moreover, StyleBreak achieves substantial improvements in attack effectiveness and efficiency across multiple attack paradigms, highlighting the urgent need for more robust alignment in LAMs.Keywords
All Related Versions
This publication has 7 references indexed in Scilit:
- Privacy dilemmas and opportunities in large language models: a brief reviewFrontiers of Computer Science, 2025
- Defending Large Language Models against Jailbreak Attacks via Semantic Smoothing2025
- Towards Holistic Evaluation of Large Audio-Language Models: A Comprehensive Survey2025
- Audio Is the Achilles’ Heel: Red Teaming Audio Large Multimodal Models2025
- AudioBench: A Universal Benchmark for Audio Large Language Models2025
- SpeechGuard: Exploring the Adversarial Robustness of Multi-modal Large Language Models2024
- Direct Preference Optimization: Your Language Model is Secretly a Reward Model2023