SilentRetrieval attack hijacks RAG systems with poisoned documents

By PulseAugur Editorial · [3 sources] · 2026-05-27 07:30

Researchers have developed "SilentRetrieval," a novel two-stage attack designed to compromise Retrieval-Augmented Generation (RAG) systems. This method uses adversarial data poisoning to inject manipulated documents that are semantically preserved and fluent, making them difficult to detect. The attack achieves high success rates in hijacking RAG outputs across various benchmarks and LLMs, even at low poisoning ratios, though defenses can mitigate its effectiveness at the cost of latency. AI

IMPACT Highlights a critical security vulnerability in RAG systems, potentially impacting the reliability and trustworthiness of AI-generated content.

RANK_REASON The cluster contains a research paper detailing a novel attack method against RAG systems.

Read on arXiv cs.IR (Information Retrieval) →

paper
safety

AI-generated summary · Google Gemini · from 3 sources. How we write summaries →

COVERAGE [3]

arXiv cs.CL TIER_1 English(EN) · Jiachen Qian · 2026-05-28 04:00

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

arXiv:2605.28074v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) mitigates LLM hallucinations but introduces a critical vulnerability: corpus integrity. We present SilentRetrieval, a two-stage data poisoning attack that hijacks RAG systems through adversaria…
arXiv cs.IR (Information Retrieval) TIER_1 English(EN) · Jiachen Qian · 2026-05-27 07:30

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

Retrieval-Augmented Generation (RAG) mitigates LLM hallucinations but introduces a critical vulnerability: corpus integrity. We present SilentRetrieval, a two-stage data poisoning attack that hijacks RAG systems through adversarially crafted yet fluent documents. Stage 1 uses Coo…
Hugging Face Daily Papers TIER_1 English(EN) · 2026-05-27 07:30

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

Retrieval-Augmented Generation (RAG) mitigates LLM hallucinations but introduces a critical vulnerability: corpus integrity. We present SilentRetrieval, a two-stage data poisoning attack that hijacks RAG systems through adversarially crafted yet fluent documents. Stage 1 uses Coo…

COVERAGE [3]

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

RELATED ENTITIES

RELATED TOPICS