PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models
Wei Zou, Runpeng Geng, Binghui Wang, Jinyuan Jia
摘要
Large language models (LLMs) have achieved remarkable success due to their exceptional generative capabilities. Despite their success, they also have inherent limitations such as a lack of up-to-date knowledge and hallucination. Retrieval-Augmented Generation (RAG) is a state-of-the-art technique to mitigate these limitations. The key idea of RAG is to ground the answer generation of an LLM on external knowledge retrieved from a knowledge database. Existing studies mainly focus on improving the accuracy or efficiency of RAG, leaving its security largely unexplored. We aim to bridge the gap in this work. We find that the knowledge database in a RAG system introduces a new and practical attack surface. Based on this attack surface, we propose PoisonedRAG, the first knowledge corruption attack to RAG, where an attacker could inject a few malicious texts into the knowledge database of a RAG system to induce an LLM to generate an attacker-chosen target answer for an attacker-chosen target question. We formulate knowledge corruption attacks as an optimization problem, whose solution is a set of malicious texts. Depending on the background knowledge (e.g., blackbox and white-box settings) of an attacker on a RAG system, we propose two solutions to solve the optimization problem, respectively. Our results show PoisonedRAG could achieve a 90% attack success rate when injecting five malicious texts for each target question into a knowledge database with millions of texts. We also evaluate several defenses and our results show they are insufficient to defend against PoisonedRAG, highlighting the need for new defenses. 1 * Equal contribution. 1 Our code is publicly available at https://github.com/sleeepeer/ PoisonedRAG Context: Sam Altman […] as the CEO of OpenAI since 2019. Question: Who is the CEO of OpenAI? Please generate a response for the question based on the context. … Tim Cook […] became the CEO of Apple in 2011. LLM Knowledge database Retriever User Question: Who is the CEO of OpenAI? Wikipedia Collect Retrieve Input Tim Cook […] became the CEO of Apple in 2011. Tim Cook […] became the CEO of Apple in 2011. Output Answer: Sam Altman User Tim Cook […] became the CEO of Apple in 2011. Tim Cook […] became the CEO of Apple in 2011. Sam Altman […] as the CEO of OpenAI since 2019.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper62
- Memory Injection Attacks on LLM Agents via Query-Only InteractionShen Dong, Shaochen Xu, Pengfei He, Yige Li 等NeurIPS 2025 · 被引用 146 次
- Your Agent May Misevolve: Emergent Risks in Self-evolving LLM AgentsShuai Shao, Qihan Ren, Dongrui Liu, Chen Qian 等ICLR 2026 · 被引用 60 次
- GraphRAG Under FireJiacheng Liang, Yuhui Wang, Changjiang Li, Tanqiu Jiang 等S&P 2026 · 被引用 31 次
- ReliabilityRAG: Effective and Provably Robust Defense for RAG-based Web-SearchZeyu Shen, Basileal Imana, Tong Wu, Chong Xiang 等NeurIPS 2025 · 被引用 26 次
- Overcoming the Retrieval Barrier: Indirect Prompt Injection in the Wild for LLM SystemsHongyan Chang, Ergute Bao, Xinjian Luo, Ting YuUSENIX Security 2026 · 被引用 24 次
它引用的顶会 Paper34
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah 等NeurIPS 2020 · 被引用 64,255 次
- Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni 等NeurIPS 2020 · 被引用 19,162 次
- Extracting Training Data from Large Language ModelsNicholas Carlini, Florian Tramèr, Eric Wallace, Matthew Jagielski 等USENIX Security 2021 · 被引用 2,866 次
- Jailbroken: How Does LLM Safety Training Fail?Alexander Wei, Nika Haghtalab, Jacob SteinhardtNeurIPS 2023 · 被引用 2,230 次
- Neural Cleanse: Identifying and Mitigating Backdoor Attacks in Neural NetworksBolun Wang, Yuanshun Yao, Shawn Shan, Huiying Li 等S&P 2019 · 被引用 1,801 次
相关 Paper
- PR-Attack: Coordinated Prompt-RAG Attacks on Retrieval-Augmented Generation in Large Language Models via Bilevel OptimizationYang Jiao, Xiaodong Wang, Kai YangSIGIR 2025 · 被引用 6 次
- KEPo: Knowledge Evolution Poison on Graph-based Retrieval-Augmented GenerationQizhi Chen, Chao Qi, Yihong Huang, Muquan Li 等WWW 2026
- Traceback of Poisoning Attacks to Retrieval-Augmented GenerationBaolei Zhang, Haoran Xin, Minghong Fang, Zhuqing Liu 等WWW 2025 · 被引用 20 次
- FlippedRAG: Black-Box Opinion Manipulation Adversarial Attacks to Retrieval-Augmented Generation ModelsZhuo Chen, Yuyang Gong, Jiawei Liu, Miaokun Chen 等CCS 2025
- Reranker Helps, but Not Enough: Towards Strong Poisoning Attacks Against Retrieval-Augmented GenerationXiaokun Yang, Jian Liang, Yesheng Liu, Xin Xiong 等ICML 2026
