New survey details RAG security and privacy risks and defenses

By PulseAugur Editorial · [1 sources] · 2026-06-24 08:08

A new survey paper published on arXiv details the security and privacy risks associated with Retrieval-Augmented Generation (RAG) systems. The paper categorizes threats across various RAG architectures, including centralized, on-device (Micro-RAG), and federated models. It outlines attack classes such as membership inference, index inference, and poisoning, while also reviewing existing defenses and highlighting the trade-offs between privacy and utility. AI

IMPACT Highlights potential vulnerabilities in RAG systems, crucial for developers building trustworthy AI applications.

RANK_REASON The cluster contains a single academic paper published on arXiv. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

paper
safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

New survey details RAG security and privacy risks and defenses

COVERAGE [1]

arXiv cs.CL TIER_1 English(EN) · Rajkumar Buyya · 2026-06-24 08:08

Security and Privacy in Retrieval-Augmented Generation: Architectures, Threats, Defenses, and Future Directions for Building Trustworthy Systems

Retrieval-Augmented Generation (RAG) has emerged as a dominant paradigm for enhancing large language models with external knowledge. By coupling retrieval mechanisms with generative models, RAG systems improve factual grounding and adaptability across domains. However, integratin…

COVERAGE [1]

Security and Privacy in Retrieval-Augmented Generation: Architectures, Threats, Defenses, and Future Directions for Building Trustworthy Systems

RELATED ENTITIES

RELATED TOPICS