@inproceedings{wang-etal-2026-hearsay,
    title = "{H}ear{S}ay Benchmark: Do Audio {LLM}s Leak What They Hear?",
    author = "Wang, Jin  and
      Luo, Kaiwen  and
      Lin, Liang  and
      Wang, Weiliu  and
      Chen, Yitian  and
      Aloqaily, Moayad  and
      Tang, Xuehai  and
      Zhou, Zhenhong  and
      Wang, Kun  and
      Sun, Li  and
      Wen, Qingsong",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.964/",
    pages = "19312--19331",
    ISBN = "979-8-89176-395-1",
    abstract = "While Audio Large Language Models (ALLMs) have achieved remarkable progress in understanding and generation, their potential privacy implications remain largely unexplored. This paper takes the first step to investigate whether ALLMs inadvertently leak user privacy solely through acoustic voiceprints and introduces $\textit{HearSay}$, a comprehensive benchmark constructed from over 22,000 real-world audio clips. To ensure data quality, the benchmark is meticulously curated through a rigorous pipeline involving automated profiling and human verification, guaranteeing that all privacy labels are grounded in factual records. Extensive experiments on $\textit{HearSay}$ yield three critical findings:$\textbf{Significant Privacy Leakage}$: ALLMs inherently extract private attributes from voiceprints, reaching 92.89{\%} accuracy on gender and effectively profiling social attributes.$\textbf{Insufficient Safety Mechanisms}$: Alarmingly, existing safeguards are severely inadequate; most models fail to refuse privacy-intruding requests, exhibiting near-zero refusal rates for physiological traits.$\textbf{Reasoning Amplifies Risk}$: Chain-of-Thought (CoT) reasoning exacerbates privacy risks in capable models by uncovering deeper acoustic correlations.These findings expose critical vulnerabilities in ALLMs, underscoring the urgent need for targeted privacy alignment.The codes and dataset are available at https://github.com/JinWang79/HearSay{\_}Benchmark"
}