AudioGuard: Toward Comprehensive Audio Safety Protection Across Diverse Threat Models

• URL: http://arxiv.org/abs/2604.08867v1
• Date: Fri, 10 Apr 2026 02:02:12 GMT
• Title: AudioGuard: Toward Comprehensive Audio Safety Protection Across Diverse Threat Models
• Abstract summary: Real-world risks can hinge on audio-native harmful sound events, speaker attributes, impersonation/voice-cloning misuse.<n>AudioGuard is a unified guardrail consisting of 1) SoundGuard for waveform-level audio-native detection and 2) ContentGuard for policy-grounded semantic protection.
• Score: 17.541986184072773
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Audio has rapidly become a primary interface for foundation models, powering real-time voice assistants. Ensuring safety in audio systems is inherently more complex than just "unsafe text spoken aloud": real-world risks can hinge on audio-native harmful sound events, speaker attributes (e.g., child voice), impersonation/voice-cloning misuse, and voice-content compositional harms, such as child voice plus sexual content. The nature of audio makes it challenging to develop comprehensive benchmarks or guardrails against this unique risk landscape. To close this gap, we conduct large-scale red teaming on audio systems, systematically uncover vulnerabilities in audio, and develop a comprehensive, policy-grounded audio risk taxonomy and AudioSafetyBench, the first policy-based audio safety benchmark across diverse threat models. AudioSafetyBench supports diverse languages, suspicious voices (e.g., celebrity/impersonation and child voice), risky voice-content combinations, and non-speech sound events. To defend against these threats, we propose AudioGuard, a unified guardrail consisting of 1) SoundGuard for waveform-level audio-native detection and 2) ContentGuard for policy-grounded semantic protection. Extensive experiments on AudioSafetyBench and four complementary benchmarks show that AudioGuard consistently improves guardrail accuracy over strong audio-LLM-based baselines with substantially lower latency.

Related papers

• UniAudio-Token: Empowering Semantic Speech Tokenizers with General Audio Perception [53.41256816230278]
  UniAudio-Token is a framework that empowers semantic tokenizers with general audio perception without compromising speech ability.<n>SAP provide structured supervision by audio into linguistic content, vocal attributes, and auditory-scene primitives.<n>SAE introduces a content-aware gating mechanism that adaptively restores fine-grained acoustic details from shallow layers.
  arXiv  Detail & Related papers  (2026-05-29T16:36:21Z)
• Hijacking Large Audio-Language Models via Context-Agnostic and Imperceptible Auditory Prompt Injection [22.306688903148046]
  Large audio-language models (LALMs) power intelligent voice interactions by tightly integrating audio and text.<n>We propose textitAudioHijack, a framework that generates context-agnostic and imperceptible adversarial audio to hijack LALMs.<n>Experiments on 13 state-of-the-art LALMs show consistent hijacking across 6 categories misbehavior.
  arXiv  Detail & Related papers  (2026-04-16T04:22:11Z)
• VoxSafeBench: Not Just What Is Said, but Who, How, and Where [30.964672315074413]
  We introduce VoxSafeBench, among the first benchmarks to jointly evaluate social alignment in speech language models.<n>Tier 1 evaluates content-centric risks using matched text and audio inputs.<n>Tier 2 targets audio-conditioned risks in which the transcript is benign but the appropriate response hinges on the speaker.<n>We find that safeguards appearing robust on text often degrade in speech.
  arXiv  Detail & Related papers  (2026-04-16T02:24:59Z)
• Now You Hear Me: Audio Narrative Attacks Against Large Audio-Language Models [26.648297855855432]
  We design a text-to-audio jailbreak that embeds disallowed directives within a narrative-style audio stream.<n>The attack exploits structural and acoustic properties, thereby circumventing safety mechanisms primarily calibrated for text.<n>Results highlight the need for safety frameworks that jointly reason over linguistic and paralinguistic representations.
  arXiv  Detail & Related papers  (2026-01-30T18:23:02Z)
• Synthetic Voices, Real Threats: Evaluating Large Text-to-Speech Models in Generating Harmful Audio [63.18443674004945]
  This work explores a content-centric threat: exploiting TTS systems to produce speech containing harmful content.<n>We present HARMGEN, a suite of five attacks organized into two families that address these challenges.
  arXiv  Detail & Related papers  (2025-11-14T03:00:04Z)
• Speech-Audio Compositional Attacks on Multimodal LLMs and Their Mitigation with SALMONN-Guard [37.736305528135944]
  We introduce SACRED-Bench to evaluate robustness of large language models (LLMs) under complex audio-based attacks.<n>We propose SALMONN-Guard, a safeguard LLM that jointly inspects speech, audio, and text for safety judgments.
  arXiv  Detail & Related papers  (2025-11-13T11:50:54Z)
• Step-Audio 2 Technical Report [120.58375054866815]
  Step-Audio 2 is an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation.<n>By integrating a latent audio encoder and reasoning-centric reinforcement learning (RL), Step-Audio 2 achieves promising performance in automatic speech recognition (ASR) and audio understanding.
  arXiv  Detail & Related papers  (2025-07-22T14:23:55Z)
• AudioTrust: Benchmarking the Multifaceted Trustworthiness of Audio Large Language Models [125.44235244336623]
  Audio Large Language Models (ALLMs) have gained widespread adoption, yet their trustworthiness remains underexplored.<n>We identify significant trustworthiness risks in ALLMs arising from non-semantic acoustic cues, including timbre, accent, and background noise.<n>We propose AudioTrust, a framework for systematic evaluation of ALLM trustworthiness across audio-specific risks.
  arXiv  Detail & Related papers  (2025-05-22T04:27:46Z)
• Audio Is the Achilles' Heel: Red Teaming Audio Large Multimodal Models [50.89022445197919]
  We show that open-source audio LMMs suffer an average attack success rate of 69.14% on harmful audio questions. Our speech-specific jailbreaks on Gemini-1.5-Pro achieve an attack success rate of 70.67% on the harmful query benchmark.
  arXiv  Detail & Related papers  (2024-10-31T12:11:17Z)
• Audio-Agent: Leveraging LLMs For Audio Generation, Editing and Composition [72.22243595269389]
  We introduce Audio-Agent, a framework for audio generation, editing and composition based on text or video inputs.<n>In our method, we utilize a pre-trained TTA diffusion network as the audio generation agent to work in tandem with GPT-4.<n>For video-to-audio (VTA) tasks, most existing methods require training a timestamp detector to synchronize video events with the generated audio.
  arXiv  Detail & Related papers  (2024-10-04T11:40:53Z)
• WaveGuard: Understanding and Mitigating Audio Adversarial Examples [12.010555227327743]
  We introduce WaveGuard: a framework for detecting adversarial inputs crafted to attack ASR systems. Our framework incorporates audio transformation functions and analyses the ASR transcriptions of the original and transformed audio to detect adversarial inputs.
  arXiv  Detail & Related papers  (2021-03-04T21:44:37Z)

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.