Yibo Zhang

The widespread application of Large Speech Models (LSMs) has made their security risks increasingly prominent. Traditional speech adversarial attack methods face challenges in balancing effectiveness and stealth. This paper proposes Evolutionary Noise Jailbreak (ENJ), which utilizes a genetic algorithm to transform environmental noise from a passive interference into an actively optimizable attack carrier for jailbreaking LSMs. Through operations such as population initialization, crossover fusion, and probabilistic mutation, this method iteratively evolves a series of audio samples that fuse malicious instructions with background noise. These samples sound like harmless noise to humans but can induce the model to parse and execute harmful commands. Extensive experiments on multiple mainstream speech models show that ENJ's attack effectiveness is significantly superior to existing baseline methods. This research reveals the dual role of noise in speech security and provides new critical insights for model security defense in complex acoustic environments.

As Audio Large Language Models (ALLMs) emerge as powerful tools for speech processing, their safety implications demand urgent attention. While considerable research has explored textual and vision safety, audio's distinct characteristics present significant challenges. This paper first investigates: Is ALLM vulnerable to backdoor attacks exploiting acoustic triggers? In response to this issue, we introduce Hidden in the Noise (HIN), a novel backdoor attack framework designed to exploit subtle, audio-specific features. HIN applies acoustic modifications to raw audio waveforms, such as alterations to temporal dynamics and strategic injection of spectrally tailored noise. These changes introduce consistent patterns that an ALLM's acoustic feature encoder captures, embedding robust triggers within the audio stream. To evaluate ALLM robustness against audio-feature-based triggers, we develop the AudioSafe benchmark, assessing nine distinct risk types. Extensive experiments on AudioSafe and three established safety datasets reveal critical vulnerabilities in existing ALLMs: (I) audio features like environment noise and speech rate variations achieve over 90% average attack success rate. (II) ALLMs exhibit significant sensitivity differences across acoustic features, particularly showing minimal response to volume as a trigger, and (III) poisoned sample inclusion causes only marginal loss curve fluctuations, highlighting the attack's stealth.