Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models

TL;DR

This paper introduces AJailBench, a comprehensive benchmark for evaluating jailbreak vulnerabilities in Large Audio Language Models, highlighting their susceptibility to adversarial audio prompts and proposing methods to generate more effective attacks.

Contribution

The paper presents the first systematic benchmark for LAM jailbreak evaluation and introduces a novel adversarial audio generation toolkit to improve attack realism and effectiveness.

Findings

None of the evaluated LAMs are consistently robust against attacks.

Small, semantically preserved perturbations can significantly compromise model safety.

The proposed adversarial toolkit enhances attack effectiveness by optimizing subtle distortions.

Abstract

The rise of Large Audio Language Models (LAMs) brings both potential and risks, as their audio outputs may contain harmful or unethical content. However, current research lacks a systematic, quantitative evaluation of LAM safety especially against jailbreak attacks, which are challenging due to the temporal and semantic nature of speech. To bridge this gap, we introduce AJailBench, the first benchmark specifically designed to evaluate jailbreak vulnerabilities in LAMs. We begin by constructing AJailBench-Base, a dataset of 1,495 adversarial audio prompts spanning 10 policy-violating categories, converted from textual jailbreak attacks using realistic text to speech synthesis. Using this dataset, we evaluate several state-of-the-art LAMs and reveal that none exhibit consistent robustness across attacks. To further strengthen jailbreak testing and simulate more realistic attack conditions, we propose a method to generate dynamic adversarial variants. Our Audio Perturbation Toolkit (APT) applies targeted distortions across time, frequency, and amplitude domains. To preserve the original jailbreak intent, we enforce a semantic consistency constraint and employ Bayesian optimization to efficiently search for perturbations that are both subtle and highly effective. This results in AJailBench-APT, an extended dataset of optimized adversarial audio samples. Our findings demonstrate that even small, semantically preserved perturbations can significantly reduce the safety performance of leading LAMs, underscoring the need for more robust and semantically aware defense mechanisms.