AI assistants can be hijacked and manipulated by inaudible sounds
A recent study reveals vulnerabilities in large audio-language models (LALMs) that can be exploited through imperceptible auditory prompt injections. Researchers developed a framework called AudioHijack, which can manipulate these models to perform unauthorized actions. The findings highlight the urgent need for improved security measures in voice interaction technologies.
- ▪The study exposes critical vulnerabilities in large audio-language models (LALMs).
- ▪AudioHijack is a framework that generates imperceptible adversarial audio to hijack LALMs.
- ▪Experiments showed success rates of 79%-96% in manipulating LALMs across various user contexts.
Hacker News (Newest) files mainly under programming. We currently carry 5,306 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | arXiv.org |
| Canonical URL | https://arxiv.org/abs/2604.14604 |
| Publication time | Sat, 23 May 2026 13:28:25 +0000 |
| Retrieval time | 2026-05-23T13:37:26.771Z |
| Last seen | 2026-05-23T13:37:26.771Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | MpJ59P9GV7nn |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Cryptography and Security arXiv:2604.14604 (cs) [Submitted on 16 Apr 2026] Title:Hijacking Large Audio-Language Models via Context-Agnostic and Imperceptible Auditory Prompt Injection Authors:Meng Chen, Kun Wang, Li Lu, Jiaheng Zhang, Tianwei Zhang View a PDF of the paper titled Hijacking Large Audio-Language Models via Context-Agnostic and Imperceptible Auditory Prompt Injection, by Meng Chen and 4 other authors View PDF HTML (experimental) Abstract:Modern Large audio-language models (LALMs) power intelligent voice interactions by tightly integrating audio and text. This integration, however, expands the attack surface beyond text and introduces vulnerabilities in the continuous, high-dimensional audio channel.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv.org.
Discussion
0 commentsMore from arXiv.org
