RAG or retrieval-augmented generation is a technique in which an AI system retrieves relevant information from external sources and uses it as context when generating an answer. The retrieved information can be current but does not have to be.
