Universal Acoustic Adversarial Attacks for Flexible Control of Speech-LLMs

Rao Ma, Mengjie Qian, Vyas Raina, Mark Gales, Kate Knill


Abstract
The combination of pre-trained speech encoders with large language models has enabled the development of speech LLMs that can handle a wide range of spoken language processing tasks. While these models are powerful and flexible, this very flexibility may make them more vulnerable to adversarial attacks. To examine the extent of this problem, in this work we investigate universal acoustic adversarial attacks on speech LLMs. Here a fixed, universal, adversarial audio segment is prepended to the original input audio. We initially investigate attacks that cause the model to either produce no output or to perform a modified task overriding the original prompt. We then extend the nature of the attack to be selective so that it activates only when specific input attributes, such as a speaker gender or spoken language, are present. Inputs without the targeted attribute should be unaffected, allowing fine-grained control over the model outputs. Our findings reveal critical vulnerabilities in Qwen2-Audio and Granite-Speech and suggest that similar speech LLMs may be susceptible to universal adversarial attacks. This highlights the need for more robust training strategies and improved resistance to adversarial attacks.
Anthology ID:
2025.findings-emnlp.990
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2025
Month:
November
Year:
2025
Address:
Suzhou, China
Editors:
Christos Christodoulopoulos, Tanmoy Chakraborty, Carolyn Rose, Violet Peng
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
18248–18262
Language:
URL:
https://aclanthology.org/2025.findings-emnlp.990/
DOI:
Bibkey:
Cite (ACL):
Rao Ma, Mengjie Qian, Vyas Raina, Mark Gales, and Kate Knill. 2025. Universal Acoustic Adversarial Attacks for Flexible Control of Speech-LLMs. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 18248–18262, Suzhou, China. Association for Computational Linguistics.
Cite (Informal):
Universal Acoustic Adversarial Attacks for Flexible Control of Speech-LLMs (Ma et al., Findings 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.findings-emnlp.990.pdf
Checklist:
 2025.findings-emnlp.990.checklist.pdf