Search papers, labs, and topics across Lattice.
This survey provides a comprehensive overview of the robustness challenges associated with Retrieval-Augmented Generation (RAG) systems, focusing on the specific vulnerabilities introduced at various stages of the pipeline. It categorizes attacks based on their objectives鈥攁ccuracy, privacy, and fairness鈥攁nd formalizes threat models for the corpus, retriever, and generator components. The findings highlight the need for stage-specific defenses and propose a structured approach to evaluate and enhance the security of RAG systems against emerging threats.
RAG systems are not just enhanced by external knowledge; they are also vulnerable to targeted attacks that can compromise accuracy, privacy, and fairness.
Retrieval-Augmented Generation (RAG) enhances large language models by grounding outputs in external knowledge, improving factuality and reducing hallucinations. At the same time, the retrieval-augmented pipeline introduces new robustness and security risks, including corpus poisoning, backdoor attacks, privacy leakage, and fairness violations. Despite rapid progress in this area, existing surveys remain limited in their treatment of attacker objectives, threat models, and stage-specific defenses across the full RAG pipeline. This survey presents a unified and pipeline-aware overview of RAG robustness. We formalize threat models over the corpus, retriever, and generator, and organize attacks into three main objectives: accuracy, privacy, and fairness. We further review defenses from a pipeline-aware perspective, covering the retrieval, rerank, generation, and traceback stages. In addition, we summarize robustness benchmarks and explainability methods for more deeply evaluating and explaining RAG robustness.
