HearSay Benchmark: Do Audio LLMs Leak What They Hear?

• URL: http://arxiv.org/abs/2601.03783v1
• Date: Wed, 07 Jan 2026 10:33:44 GMT
• Title: HearSay Benchmark: Do Audio LLMs Leak What They Hear?
• Authors: Jin Wang, Liang Lin, Kaiwen Luo, Weiliu Wang, Yitian Chen, Moayad Aloqaily, Xuehai Tang, Zhenhong Zhou, Kun Wang, Li Sun, Qingsong Wen,
• Abstract summary: This paper investigates whether Audio Large Language Models inadvertently leak user privacy solely through acoustic voiceprints.<n>$textitHearSay$ is a comprehensive benchmark constructed from over 22,000 real-world audio clips.<n>Experiments on $textbfSignificant Privacy Leakage$ yield three critical findings.
• Score: 71.05839007164776
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• 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

Related papers

• LoRA and Privacy: When Random Projections Help (and When They Don't) [55.65932772290123]
  We introduce the (Wishart) projection mechanism, a randomized map of the form $S mapsto M f(S)$ with $M sim W_d (1/r I_d, r)$ and study its differential privacy properties.<n>For vector-valued queries $f$, we prove non-asymptotic DP guarantees without any additive noise, showing that Wishart randomness alone can suffice.<n>For matrix-valued queries, however, we establish a sharp negative result: in the noise-free setting, the mechanism is not DP, and we demonstrate its vulnerability.
  arXiv  Detail & Related papers  (2026-01-29T13:43:37Z)
• PrivacyBench: A Conversational Benchmark for Evaluating Privacy in Personalized AI [8.799432439533211]
  AI agents rely on access to a user's digital footprint, which often includes sensitive data from private emails, chats and purchase histories.<n>This access creates a fundamental societal and privacy risk.<n>We introduce PrivacyBench, a benchmark with socially grounded datasets containing embedded secrets.
  arXiv  Detail & Related papers  (2025-12-31T13:16:45Z)
• High-Dimensional Asymptotics of Differentially Private PCA [4.168157981135696]
  In differential privacy, statistics of a sensitive dataset are privatized by introducing random noise.<n>It remains unclear if such high noise levels are truly necessary or a limitation of the proof techniques.<n>This paper explores whether we can obtain sharp privacy characterizations that identify the smallest noise level required to achieve a target privacy level.
  arXiv  Detail & Related papers  (2025-11-10T16:17:16Z)
• VoxGuard: Evaluating User and Attribute Privacy in Speech via Membership Inference Attacks [51.68795949691009]
  We introduce VoxGuard, a framework grounded in differential privacy and membership inference.<n>For attributes, we show that simple transparent attacks recover gender and accent with near-perfect accuracy even after anonymization.<n>Our results demonstrate that EER substantially underestimates leakage, highlighting the need for low-FPR evaluation.
  arXiv  Detail & Related papers  (2025-09-22T20:57:48Z)
• Privacy-Aware Decoding: Mitigating Privacy Leakage of Large Language Models in Retrieval-Augmented Generation [26.573578326262307]
  Privacy-Aware Decoding (PAD) is a lightweight, inference-time defense that adaptively injects calibrated Gaussian noise into token logits during generation.<n>PAD integrates confidence-based screening to selectively protect high-risk tokens, efficient sensitivity estimation to minimize unnecessary noise, and context-aware noise calibration to balance privacy with generation quality.<n>Our work takes an important step toward mitigating privacy risks in RAG via decoding strategies, paving the way for universal and scalable privacy solutions in sensitive domains.
  arXiv  Detail & Related papers  (2025-08-05T05:22:13Z)
• DP-Fusion: Token-Level Differentially Private Inference for Large Language Models [51.71591819896191]
  Large language models (LLMs) do not preserve privacy at inference-time.<n> DP-Fusion provably bounds the influence a set of tokens in the context can have on the LLM's output.<n>We show that our method creates token-level provably privatized documents with substantially improved theoretical and empirical privacy.
  arXiv  Detail & Related papers  (2025-07-06T20:49:39Z)
• InvisibleInk: High-Utility and Low-Cost Text Generation with Differential Privacy [7.006059299522521]
  InvisibleInk is a scalable long-form text generation framework satisfying rigorous differential privacy guarantees.<n>We reduce the privacy cost by isolating and clipping only the sensitive information in the model logits.<n>We improve text quality by sampling from a small superset of the top-$k$ private tokens.
  arXiv  Detail & Related papers  (2025-06-30T18:00:41Z)
• Privacy-Preserving Retrieval-Augmented Generation with Differential Privacy [25.896416088293908]
  retrieval-augmented generation (RAG) is particularly effective in assisting large language models (LLMs)<n>RAG outputs risk leaking sensitive information from the external data source.<n>We propose an algorithm that smartly spends privacy budget only for the tokens that require the sensitive information.
  arXiv  Detail & Related papers  (2024-12-06T01:20:16Z)
• Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theory [82.7042006247124]
  We show that even the most capable AI models reveal private information in contexts that humans would not, 39% and 57% of the time, respectively. Our work underscores the immediate need to explore novel inference-time privacy-preserving approaches, based on reasoning and theory of mind.
  arXiv  Detail & Related papers  (2023-10-27T04:15:30Z)
• Adaptive Privacy Composition for Accuracy-first Mechanisms [55.53725113597539]
  Noise reduction mechanisms produce increasingly accurate answers. Analysts only pay the privacy cost of the least noisy or most accurate answer released. There has yet to be any study on how ex-post private mechanisms compose. We develop privacy filters that allow an analyst to adaptively switch between differentially private and ex-post private mechanisms.
  arXiv  Detail & Related papers  (2023-06-24T00:33:34Z)
• Privacy Implications of Retrieval-Based Language Models [26.87950501433784]
  We present the first study of privacy risks in retrieval-based LMs, particularly $k$NN-LMs. We find that $k$NN-LMs are more susceptible to leaking private information from their private datastore than parametric models.
  arXiv  Detail & Related papers  (2023-05-24T08:37:27Z)
• An Experimental Study on Private Aggregation of Teacher Ensemble Learning for End-to-End Speech Recognition [51.232523987916636]
  Differential privacy (DP) is one data protection avenue to safeguard user information used for training deep models by imposing noisy distortion on privacy data. In this work, we extend PATE learning to work with dynamic patterns, namely speech, and perform one very first experimental study on ASR to avoid acoustic data leakage.
  arXiv  Detail & Related papers  (2022-10-11T16:55:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.