This study assesses the vulnerability of Retrieval-Augmented Generation systems to minor textual perturbations. A new attack method, the Genetic Attack on RAG (GARAG), has been developed to disclose these weaknesses across various components of the RAG system, thus testing its overall robustness. ### Summary: Risk Factors: - Minor textual errors can lead to significant issues in the functionality of RAG systems. - Evaluates the impact across different system components. ### Opinion: The insights from this research could be crucial in fortifying the robustness of RAG systems, making them more resilient against subtle but potentially disruptive errors. Such advancements are vital for the ongoing development and security of AI technologies.