Study: RAG-based LLMs less safe than non-RAG

By Anthony Malakian, waterstechnology, 29 Apr 2025 URL: https://www.waterstechnology.com/emerging-technologies/7952391/study-rag-based-llms-less-safe-than-non-rag

In the world of large language models, the retrieval-augmented generation (RAG) framework has gained favor as companies have looked to develop LLMs. RAG approaches grew in popularity about a year ago as the framework was viewed as a way to reduce hallucinations related to generative AI. IBM wrote in August 2023 that RAG LLMs have access to the most current, reliable facts. Broadly, RAG-based LLMs have been deemed safer from the danger of answering “harmful questions” than non-RAG LLMs—but a new…