The Hidden Threat in Plain Text: Attacking RAG Data Loaders
Alberto Castagnaro, Umberto Salviati, Mauro Conti, Luca Pajola, Simeone Pizzi · 2025
Large Language Models (LLMs) have transformed human–machine interaction since ChatGPT’s 2022 debut, with Retrieval-Augmented Generation (RAG) emerging as a key framework that enhances LLM outputs by integrating external knowledge. However, RAG’s reliance on ingesting external documents introduces new vulnerabilities. This paper exposes a critical security gap at the data loading stage, where malicious actors can stealthily corrupt RAG pipelines by exploiting document ingestion.