# Audio Hallucination Attacks: Probing the Reliability of Large Audio Language Models

arXiv: https://arxiv.org/abs/2603.29263 · HTML: https://mlsanity.com/paper/2603.29263/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 31 Mar 2026.

## Built on by (most cited first, 1 of 1)

- [SteerDuplex: Steerable Duplex Speech Dialogue Models](https://mlsanity.com/paper/2609.12623/) (2026)

## Abstract (arXiv)

Large Audio Language Models (LALMs) achieve strong performance on audio-language tasks; however, their reliability in real-world settings remains underexplored. We introduce Audio Hallucination Attacks (AHA), an attack suite called AHA-Eval, comprising 6.5K QA pairs designed to test whether LALMs genuinely ground their responses in the audio input. AHA targets two attack surfaces: (i) query-based attacks, which exploit question structure to induce hallucinations about absent sounds, and (ii) audio-based attacks, which inject synthetic speech describing non-existent events into the audio stream. Evaluating state-of-the-art LALMs, including Audio Flamingo 3 and Gemini 3 Pro, we observe high attack success rates of 95.35% and 79.65%, respectively, revealing a reliability gap that is hidden by standard benchmark performance. To mitigate this, we propose a 120K QA post-alignment dataset, AHA-Guard, which successfully reduces attack success rates by up to 49%.